thesis.lyx 97 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768697071727374757677787980818283848586878889909192939495969798991001011021031041051061071081091101111121131141151161171181191201211221231241251261271281291301311321331341351361371381391401411421431441451461471481491501511521531541551561571581591601611621631641651661671681691701711721731741751761771781791801811821831841851861871881891901911921931941951961971981992002012022032042052062072082092102112122132142152162172182192202212222232242252262272282292302312322332342352362372382392402412422432442452462472482492502512522532542552562572582592602612622632642652662672682692702712722732742752762772782792802812822832842852862872882892902912922932942952962972982993003013023033043053063073083093103113123133143153163173183193203213223233243253263273283293303313323333343353363373383393403413423433443453463473483493503513523533543553563573583593603613623633643653663673683693703713723733743753763773783793803813823833843853863873883893903913923933943953963973983994004014024034044054064074084094104114124134144154164174184194204214224234244254264274284294304314324334344354364374384394404414424434444454464474484494504514524534544554564574584594604614624634644654664674684694704714724734744754764774784794804814824834844854864874884894904914924934944954964974984995005015025035045055065075085095105115125135145155165175185195205215225235245255265275285295305315325335345355365375385395405415425435445455465475485495505515525535545555565575585595605615625635645655665675685695705715725735745755765775785795805815825835845855865875885895905915925935945955965975985996006016026036046056066076086096106116126136146156166176186196206216226236246256266276286296306316326336346356366376386396406416426436446456466476486496506516526536546556566576586596606616626636646656666676686696706716726736746756766776786796806816826836846856866876886896906916926936946956966976986997007017027037047057067077087097107117127137147157167177187197207217227237247257267277287297307317327337347357367377387397407417427437447457467477487497507517527537547557567577587597607617627637647657667677687697707717727737747757767777787797807817827837847857867877887897907917927937947957967977987998008018028038048058068078088098108118128138148158168178188198208218228238248258268278288298308318328338348358368378388398408418428438448458468478488498508518528538548558568578588598608618628638648658668678688698708718728738748758768778788798808818828838848858868878888898908918928938948958968978988999009019029039049059069079089099109119129139149159169179189199209219229239249259269279289299309319329339349359369379389399409419429439449459469479489499509519529539549559569579589599609619629639649659669679689699709719729739749759769779789799809819829839849859869879889899909919929939949959969979989991000100110021003100410051006100710081009101010111012101310141015101610171018101910201021102210231024102510261027102810291030103110321033103410351036103710381039104010411042104310441045104610471048104910501051105210531054105510561057105810591060106110621063106410651066106710681069107010711072107310741075107610771078107910801081108210831084108510861087108810891090109110921093109410951096109710981099110011011102110311041105110611071108110911101111111211131114111511161117111811191120112111221123112411251126112711281129113011311132113311341135113611371138113911401141114211431144114511461147114811491150115111521153115411551156115711581159116011611162116311641165116611671168116911701171117211731174117511761177117811791180118111821183118411851186118711881189119011911192119311941195119611971198119912001201120212031204120512061207120812091210121112121213121412151216121712181219122012211222122312241225122612271228122912301231123212331234123512361237123812391240124112421243124412451246124712481249125012511252125312541255125612571258125912601261126212631264126512661267126812691270127112721273127412751276127712781279128012811282128312841285128612871288128912901291129212931294129512961297129812991300130113021303130413051306130713081309131013111312131313141315131613171318131913201321132213231324132513261327132813291330133113321333133413351336133713381339134013411342134313441345134613471348134913501351135213531354135513561357135813591360136113621363136413651366136713681369137013711372137313741375137613771378137913801381138213831384138513861387138813891390139113921393139413951396139713981399140014011402140314041405140614071408140914101411141214131414141514161417141814191420142114221423142414251426142714281429143014311432143314341435143614371438143914401441144214431444144514461447144814491450145114521453145414551456145714581459146014611462146314641465146614671468146914701471147214731474147514761477147814791480148114821483148414851486148714881489149014911492149314941495149614971498149915001501150215031504150515061507150815091510151115121513151415151516151715181519152015211522152315241525152615271528152915301531153215331534153515361537153815391540154115421543154415451546154715481549155015511552155315541555155615571558155915601561156215631564156515661567156815691570157115721573157415751576157715781579158015811582158315841585158615871588158915901591159215931594159515961597159815991600160116021603160416051606160716081609161016111612161316141615161616171618161916201621162216231624162516261627162816291630163116321633163416351636163716381639164016411642164316441645164616471648164916501651165216531654165516561657165816591660166116621663166416651666166716681669167016711672167316741675167616771678167916801681168216831684168516861687168816891690169116921693169416951696169716981699170017011702170317041705170617071708170917101711171217131714171517161717171817191720172117221723172417251726172717281729173017311732173317341735173617371738173917401741174217431744174517461747174817491750175117521753175417551756175717581759176017611762176317641765176617671768176917701771177217731774177517761777177817791780178117821783178417851786178717881789179017911792179317941795179617971798179918001801180218031804180518061807180818091810181118121813181418151816181718181819182018211822182318241825182618271828182918301831183218331834183518361837183818391840184118421843184418451846184718481849185018511852185318541855185618571858185918601861186218631864186518661867186818691870187118721873187418751876187718781879188018811882188318841885188618871888188918901891189218931894189518961897189818991900190119021903190419051906190719081909191019111912191319141915191619171918191919201921192219231924192519261927192819291930193119321933193419351936193719381939194019411942194319441945194619471948194919501951195219531954195519561957195819591960196119621963196419651966196719681969197019711972197319741975197619771978197919801981198219831984198519861987198819891990199119921993199419951996199719981999200020012002200320042005200620072008200920102011201220132014201520162017201820192020202120222023202420252026202720282029203020312032203320342035203620372038203920402041204220432044204520462047204820492050205120522053205420552056205720582059206020612062206320642065206620672068206920702071207220732074207520762077207820792080208120822083208420852086208720882089209020912092209320942095209620972098209921002101210221032104210521062107210821092110211121122113211421152116211721182119212021212122212321242125212621272128212921302131213221332134213521362137213821392140214121422143214421452146214721482149215021512152215321542155215621572158215921602161216221632164216521662167216821692170217121722173217421752176217721782179218021812182218321842185218621872188218921902191219221932194219521962197219821992200220122022203220422052206220722082209221022112212221322142215221622172218221922202221222222232224222522262227222822292230223122322233223422352236223722382239224022412242224322442245224622472248224922502251225222532254225522562257225822592260226122622263226422652266226722682269227022712272227322742275227622772278227922802281228222832284228522862287228822892290229122922293229422952296229722982299230023012302230323042305230623072308230923102311231223132314231523162317231823192320232123222323232423252326232723282329233023312332233323342335233623372338233923402341234223432344234523462347234823492350235123522353235423552356235723582359236023612362236323642365236623672368236923702371237223732374237523762377237823792380238123822383238423852386238723882389239023912392239323942395239623972398239924002401240224032404240524062407240824092410241124122413241424152416241724182419242024212422242324242425242624272428242924302431243224332434243524362437243824392440244124422443244424452446244724482449245024512452245324542455245624572458245924602461246224632464246524662467246824692470247124722473247424752476247724782479248024812482248324842485248624872488248924902491249224932494249524962497249824992500250125022503250425052506250725082509251025112512251325142515251625172518251925202521252225232524252525262527252825292530253125322533253425352536253725382539254025412542254325442545254625472548254925502551255225532554255525562557255825592560256125622563256425652566256725682569257025712572257325742575257625772578257925802581258225832584258525862587258825892590259125922593259425952596259725982599260026012602260326042605260626072608260926102611261226132614261526162617261826192620262126222623262426252626262726282629263026312632263326342635263626372638263926402641264226432644264526462647264826492650265126522653265426552656265726582659266026612662266326642665266626672668266926702671267226732674267526762677267826792680268126822683268426852686268726882689269026912692269326942695269626972698269927002701270227032704270527062707270827092710271127122713271427152716271727182719272027212722272327242725272627272728272927302731273227332734273527362737273827392740274127422743274427452746274727482749275027512752275327542755275627572758275927602761276227632764276527662767276827692770277127722773277427752776277727782779278027812782278327842785278627872788278927902791279227932794279527962797279827992800280128022803280428052806280728082809281028112812281328142815281628172818281928202821282228232824282528262827282828292830283128322833283428352836283728382839284028412842284328442845284628472848284928502851285228532854285528562857285828592860286128622863286428652866286728682869287028712872287328742875287628772878287928802881288228832884288528862887288828892890289128922893289428952896289728982899290029012902290329042905290629072908290929102911291229132914291529162917291829192920292129222923292429252926292729282929293029312932293329342935293629372938293929402941294229432944294529462947294829492950295129522953295429552956295729582959296029612962296329642965296629672968296929702971297229732974297529762977297829792980298129822983298429852986298729882989299029912992299329942995299629972998299930003001300230033004300530063007300830093010301130123013301430153016301730183019302030213022302330243025302630273028302930303031303230333034303530363037303830393040304130423043304430453046304730483049305030513052305330543055305630573058305930603061306230633064306530663067306830693070307130723073307430753076307730783079308030813082308330843085308630873088308930903091309230933094309530963097309830993100310131023103310431053106310731083109311031113112311331143115311631173118311931203121312231233124312531263127312831293130313131323133313431353136313731383139314031413142314331443145314631473148314931503151315231533154315531563157315831593160316131623163316431653166316731683169317031713172317331743175317631773178317931803181318231833184318531863187318831893190319131923193319431953196319731983199320032013202320332043205320632073208320932103211321232133214321532163217321832193220322132223223322432253226322732283229323032313232323332343235323632373238323932403241324232433244324532463247324832493250325132523253325432553256325732583259326032613262326332643265326632673268326932703271327232733274327532763277327832793280328132823283328432853286328732883289329032913292329332943295329632973298329933003301330233033304330533063307330833093310331133123313331433153316331733183319332033213322332333243325332633273328332933303331333233333334333533363337333833393340334133423343334433453346334733483349335033513352335333543355335633573358335933603361336233633364336533663367336833693370337133723373337433753376337733783379338033813382338333843385338633873388338933903391339233933394339533963397339833993400340134023403340434053406340734083409341034113412341334143415341634173418341934203421342234233424342534263427342834293430343134323433343434353436343734383439344034413442344334443445344634473448344934503451345234533454345534563457345834593460346134623463346434653466346734683469347034713472347334743475347634773478347934803481348234833484348534863487348834893490349134923493349434953496349734983499350035013502350335043505350635073508350935103511351235133514351535163517351835193520352135223523352435253526352735283529353035313532353335343535353635373538353935403541354235433544354535463547354835493550355135523553355435553556355735583559356035613562356335643565356635673568356935703571357235733574357535763577357835793580358135823583358435853586358735883589359035913592359335943595359635973598359936003601360236033604360536063607360836093610361136123613361436153616361736183619362036213622362336243625362636273628362936303631363236333634363536363637363836393640364136423643364436453646364736483649365036513652365336543655365636573658365936603661366236633664366536663667366836693670367136723673367436753676367736783679368036813682368336843685368636873688368936903691369236933694369536963697369836993700370137023703370437053706370737083709371037113712371337143715371637173718371937203721372237233724372537263727372837293730373137323733373437353736373737383739374037413742374337443745374637473748374937503751375237533754375537563757375837593760376137623763376437653766376737683769377037713772377337743775377637773778377937803781378237833784378537863787378837893790379137923793379437953796379737983799380038013802380338043805380638073808380938103811381238133814381538163817381838193820382138223823382438253826382738283829383038313832383338343835383638373838383938403841384238433844384538463847384838493850385138523853385438553856385738583859386038613862386338643865386638673868386938703871387238733874387538763877387838793880388138823883388438853886388738883889389038913892389338943895389638973898389939003901390239033904390539063907390839093910391139123913391439153916391739183919392039213922392339243925392639273928392939303931393239333934393539363937393839393940394139423943394439453946394739483949395039513952395339543955395639573958395939603961396239633964396539663967396839693970397139723973397439753976397739783979398039813982398339843985398639873988398939903991399239933994399539963997399839994000400140024003400440054006400740084009401040114012401340144015401640174018401940204021402240234024402540264027402840294030403140324033403440354036403740384039404040414042404340444045404640474048404940504051405240534054405540564057405840594060406140624063406440654066406740684069407040714072407340744075407640774078407940804081408240834084408540864087408840894090409140924093409440954096
  1. #LyX 2.3 created this file. For more info see http://www.lyx.org/
  2. \lyxformat 544
  3. \begin_document
  4. \begin_header
  5. \save_transient_properties true
  6. \origin unavailable
  7. \textclass extbook
  8. \begin_preamble
  9. % List all used files in log output
  10. \listfiles
  11. % Add a DRAFT watermark
  12. \usepackage{draftwatermark}
  13. \SetWatermarkLightness{0.97}
  14. \SetWatermarkScale{1}
  15. % Set up required header format
  16. \usepackage{fancyhdr}
  17. \pagestyle{fancy}
  18. \renewcommand{\headrulewidth}{0pt}
  19. \rhead{}
  20. \lhead{}
  21. \rfoot{}
  22. \lfoot{}
  23. \cfoot{\thepage} % Page number bottom center
  24. % https://tex.stackexchange.com/questions/65680/automatically-bold-first-sentence-of-a-floats-caption
  25. \usepackage{xstring}
  26. \usepackage{etoolbox}
  27. \usepackage{caption}
  28. \captionsetup{labelfont=bf,tableposition=top}
  29. \makeatletter
  30. \newcommand\formatlabel[1]{%
  31. \noexpandarg
  32. \IfSubStr{#1}{.}{%
  33. \StrBefore{#1}{.}[\firstcaption]%
  34. \StrBehind{#1}{.}[\secondcaption]%
  35. \textbf{\firstcaption.} \secondcaption}{%
  36. #1}%
  37. }
  38. \patchcmd{\@caption}{#3}{\formatlabel{#3}}
  39. \makeatother
  40. \end_preamble
  41. \use_default_options true
  42. \begin_modules
  43. todonotes
  44. \end_modules
  45. \maintain_unincluded_children false
  46. \language english
  47. \language_package default
  48. \inputencoding utf8
  49. \fontencoding default
  50. \font_roman "default" "default"
  51. \font_sans "default" "default"
  52. \font_typewriter "default" "default"
  53. \font_math "auto" "auto"
  54. \font_default_family default
  55. \use_non_tex_fonts false
  56. \font_sc false
  57. \font_osf false
  58. \font_sf_scale 100 100
  59. \font_tt_scale 100 100
  60. \use_microtype false
  61. \use_dash_ligatures true
  62. \graphics default
  63. \default_output_format pdf4
  64. \output_sync 0
  65. \bibtex_command default
  66. \index_command default
  67. \paperfontsize 12
  68. \spacing double
  69. \use_hyperref true
  70. \pdf_bookmarks true
  71. \pdf_bookmarksnumbered false
  72. \pdf_bookmarksopen false
  73. \pdf_bookmarksopenlevel 1
  74. \pdf_breaklinks false
  75. \pdf_pdfborder false
  76. \pdf_colorlinks false
  77. \pdf_backref false
  78. \pdf_pdfusetitle true
  79. \papersize letterpaper
  80. \use_geometry true
  81. \use_package amsmath 1
  82. \use_package amssymb 1
  83. \use_package cancel 1
  84. \use_package esint 1
  85. \use_package mathdots 1
  86. \use_package mathtools 1
  87. \use_package mhchem 1
  88. \use_package stackrel 1
  89. \use_package stmaryrd 1
  90. \use_package undertilde 1
  91. \cite_engine basic
  92. \cite_engine_type default
  93. \biblio_style plain
  94. \use_bibtopic false
  95. \use_indices false
  96. \paperorientation portrait
  97. \suppress_date false
  98. \justification true
  99. \use_refstyle 1
  100. \use_minted 0
  101. \index Index
  102. \shortcut idx
  103. \color #008000
  104. \end_index
  105. \leftmargin 1.5in
  106. \topmargin 1in
  107. \rightmargin 1in
  108. \bottommargin 1in
  109. \secnumdepth 3
  110. \tocdepth 3
  111. \paragraph_separation indent
  112. \paragraph_indentation default
  113. \is_math_indent 0
  114. \math_numbering_side default
  115. \quotes_style english
  116. \dynamic_quotes 0
  117. \papercolumns 1
  118. \papersides 2
  119. \paperpagestyle default
  120. \tracking_changes false
  121. \output_changes false
  122. \html_math_output 0
  123. \html_css_as_file 0
  124. \html_be_strict false
  125. \end_header
  126. \begin_body
  127. \begin_layout Title
  128. Bioinformatic analysis of complex, high-throughput genomic and epigenomic
  129. data in the context of immunology and transplant rejection
  130. \end_layout
  131. \begin_layout Author
  132. A thesis presented
  133. \begin_inset Newline newline
  134. \end_inset
  135. by
  136. \begin_inset Newline newline
  137. \end_inset
  138. Ryan C.
  139. Thompson
  140. \begin_inset Newline newline
  141. \end_inset
  142. to
  143. \begin_inset Newline newline
  144. \end_inset
  145. The Scripps Research Institute Graduate Program
  146. \begin_inset Newline newline
  147. \end_inset
  148. in partial fulfillment of the requirements for the degree of
  149. \begin_inset Newline newline
  150. \end_inset
  151. Doctor of Philosophy in the subject of Biology
  152. \begin_inset Newline newline
  153. \end_inset
  154. for
  155. \begin_inset Newline newline
  156. \end_inset
  157. The Scripps Research Institute
  158. \begin_inset Newline newline
  159. \end_inset
  160. La Jolla, California
  161. \end_layout
  162. \begin_layout Date
  163. May 2019
  164. \end_layout
  165. \begin_layout Standard
  166. [Copyright notice]
  167. \end_layout
  168. \begin_layout Standard
  169. [Thesis acceptance form]
  170. \end_layout
  171. \begin_layout Standard
  172. [Dedication]
  173. \end_layout
  174. \begin_layout Standard
  175. [Acknowledgements]
  176. \end_layout
  177. \begin_layout Standard
  178. \begin_inset CommandInset toc
  179. LatexCommand tableofcontents
  180. \end_inset
  181. \end_layout
  182. \begin_layout Standard
  183. \begin_inset FloatList table
  184. \end_inset
  185. \end_layout
  186. \begin_layout Standard
  187. \begin_inset FloatList figure
  188. \end_inset
  189. \end_layout
  190. \begin_layout Standard
  191. [List of Abbreviations]
  192. \end_layout
  193. \begin_layout Standard
  194. \begin_inset Flex TODO Note (inline)
  195. status open
  196. \begin_layout Plain Layout
  197. Look into auto-generated nomenclature list: https://wiki.lyx.org/Tips/Nomenclature
  198. \end_layout
  199. \end_inset
  200. \end_layout
  201. \begin_layout List of TODOs
  202. \end_layout
  203. \begin_layout Standard
  204. [Abstract]
  205. \end_layout
  206. \begin_layout Chapter*
  207. Abstract
  208. \end_layout
  209. \begin_layout Chapter
  210. Introduction
  211. \end_layout
  212. \begin_layout Section
  213. Background & Significance
  214. \end_layout
  215. \begin_layout Subsection
  216. Biological motivation
  217. \end_layout
  218. \begin_layout Itemize
  219. Rejection is the major long-term threat to organ and tissue grafts
  220. \end_layout
  221. \begin_deeper
  222. \begin_layout Itemize
  223. Common mechanisms of rejection
  224. \end_layout
  225. \begin_layout Itemize
  226. Effective immune suppression requires monitoring for rejection and tuning
  227. \end_layout
  228. \begin_layout Itemize
  229. Current tests for rejection (tissue biopsy) are invasive and biased
  230. \end_layout
  231. \begin_layout Itemize
  232. A blood test based on microarrays would be less biased and invasive
  233. \end_layout
  234. \end_deeper
  235. \begin_layout Itemize
  236. Memory cells are resistant to immune suppression
  237. \end_layout
  238. \begin_deeper
  239. \begin_layout Itemize
  240. Mechanisms of resistance in memory cells are poorly understood
  241. \end_layout
  242. \begin_layout Itemize
  243. A better understanding of immune memory formation is needed
  244. \end_layout
  245. \end_deeper
  246. \begin_layout Itemize
  247. Mesenchymal stem cell infusion is a promising new treatment to prevent/delay
  248. rejection
  249. \end_layout
  250. \begin_deeper
  251. \begin_layout Itemize
  252. Demonstrated in mice, but not yet in primates
  253. \end_layout
  254. \begin_layout Itemize
  255. Mechanism currently unknown, but MSC are known to be immune modulatory
  256. \end_layout
  257. \end_deeper
  258. \begin_layout Subsection
  259. Overview of bioinformatic analysis methods
  260. \end_layout
  261. \begin_layout Standard
  262. An overview of all the methods used, including what problem they solve,
  263. what assumptions they make, and a basic description of how they work.
  264. \end_layout
  265. \begin_layout Itemize
  266. ChIP-seq Peak calling
  267. \end_layout
  268. \begin_deeper
  269. \begin_layout Itemize
  270. Cross-correlation analysis to determine fragment size
  271. \end_layout
  272. \begin_layout Itemize
  273. Broad vs narrow peaks
  274. \end_layout
  275. \begin_layout Itemize
  276. SICER for broad peaks
  277. \end_layout
  278. \begin_layout Itemize
  279. IDR for biologically reproducible peaks
  280. \end_layout
  281. \begin_layout Itemize
  282. csaw peak filtering guidelines for unbiased downstream analysis
  283. \end_layout
  284. \end_deeper
  285. \begin_layout Itemize
  286. Normalization is non-trivial and application-dependant
  287. \end_layout
  288. \begin_deeper
  289. \begin_layout Itemize
  290. Expression arrays: RMA & fRMA; why fRMA is needed
  291. \end_layout
  292. \begin_layout Itemize
  293. Methylation arrays: M-value transformation approximates normal data but
  294. induces heteroskedasticity
  295. \end_layout
  296. \begin_layout Itemize
  297. RNA-seq: normalize based on assumption that the average gene is not changing
  298. \end_layout
  299. \begin_layout Itemize
  300. ChIP-seq: complex with many considerations, dependent on experimental methods,
  301. biological system, and analysis goals
  302. \end_layout
  303. \end_deeper
  304. \begin_layout Itemize
  305. Limma: The standard linear modeling framework for genomics
  306. \end_layout
  307. \begin_deeper
  308. \begin_layout Itemize
  309. empirical Bayes variance modeling: limma's core feature
  310. \end_layout
  311. \begin_layout Itemize
  312. edgeR & DESeq2: Extend with negative bonomial GLM for RNA-seq and other
  313. count data
  314. \end_layout
  315. \begin_layout Itemize
  316. voom: Extend with precision weights to model mean-variance trend
  317. \end_layout
  318. \begin_layout Itemize
  319. arrayWeights and duplicateCorrelation to handle complex variance structures
  320. \end_layout
  321. \end_deeper
  322. \begin_layout Itemize
  323. sva and ComBat for batch correction
  324. \end_layout
  325. \begin_layout Itemize
  326. Factor analysis: PCA, MDS, MOFA
  327. \end_layout
  328. \begin_deeper
  329. \begin_layout Itemize
  330. Batch-corrected PCA is informative, but careful application is required
  331. to avoid bias
  332. \end_layout
  333. \end_deeper
  334. \begin_layout Itemize
  335. Gene set analysis: camera and SPIA
  336. \end_layout
  337. \begin_layout Section
  338. Innovation
  339. \end_layout
  340. \begin_layout Itemize
  341. MSC infusion to improve transplant outcomes (prevent/delay rejection)
  342. \end_layout
  343. \begin_deeper
  344. \begin_layout Itemize
  345. Characterize MSC response to interferon gamma
  346. \end_layout
  347. \begin_layout Itemize
  348. IFN-g is thought to stimulate their function
  349. \end_layout
  350. \begin_layout Itemize
  351. Test IFN-g treated MSC infusion as a therapy to delay graft rejection in
  352. cynomolgus monkeys
  353. \end_layout
  354. \begin_layout Itemize
  355. Monitor animals post-transplant using blood RNA-seq at serial time points
  356. \end_layout
  357. \end_deeper
  358. \begin_layout Itemize
  359. Investigate dynamics of histone marks in CD4 T-cell activation and memory
  360. \end_layout
  361. \begin_deeper
  362. \begin_layout Itemize
  363. Previous studies have looked at single snapshots of histone marks
  364. \end_layout
  365. \begin_layout Itemize
  366. Instead, look at changes in histone marks across activation and memory
  367. \end_layout
  368. \end_deeper
  369. \begin_layout Itemize
  370. High-throughput sequencing and microarray technologies
  371. \end_layout
  372. \begin_deeper
  373. \begin_layout Itemize
  374. Powerful methods for assaying gene expression and epigenetics across entire
  375. genomes
  376. \end_layout
  377. \begin_layout Itemize
  378. Proper analysis requires finding and exploiting systematic genome-wide trends
  379. \end_layout
  380. \end_deeper
  381. \begin_layout Chapter
  382. Reproducible genome-wide epigenetic analysis of H3K4 and H3K27 methylation
  383. in naive and memory CD4 T-cell activation
  384. \end_layout
  385. \begin_layout Standard
  386. \begin_inset Flex TODO Note (inline)
  387. status open
  388. \begin_layout Plain Layout
  389. Author list: Me, Sarah, Dan
  390. \end_layout
  391. \end_inset
  392. \end_layout
  393. \begin_layout Section
  394. Approach
  395. \end_layout
  396. \begin_layout Itemize
  397. CD4 T-cells are central to all adaptive immune responses and memory
  398. \end_layout
  399. \begin_layout Itemize
  400. H3K4 and H3K27 methylation are major epigenetic regulators of gene expression
  401. \end_layout
  402. \begin_layout Itemize
  403. Canonically, H3K4 is activating and H3K27 is inhibitory, but the reality
  404. is complex
  405. \end_layout
  406. \begin_layout Itemize
  407. Looking at these marks during CD4 activation and memory should reveal new
  408. mechanistic details
  409. \end_layout
  410. \begin_layout Itemize
  411. Test
  412. \begin_inset Quotes eld
  413. \end_inset
  414. poised promoter
  415. \begin_inset Quotes erd
  416. \end_inset
  417. hypothesis in which H3K4 and H3K27 are both methylated
  418. \end_layout
  419. \begin_layout Itemize
  420. Expand scope of analysis beyond simple promoter counts
  421. \end_layout
  422. \begin_deeper
  423. \begin_layout Itemize
  424. Analyze peaks genome-wide, including in intergenic regions
  425. \end_layout
  426. \begin_layout Itemize
  427. Analysis of coverage distribution shape within promoters, e.g.
  428. upstream vs downstream coverage
  429. \end_layout
  430. \end_deeper
  431. \begin_layout Section
  432. Methods
  433. \end_layout
  434. \begin_layout Itemize
  435. Re-analyze previously published CD4 ChIP-seq & RNA-seq data
  436. \begin_inset CommandInset citation
  437. LatexCommand cite
  438. key "LaMere2016,Lamere2017"
  439. literal "true"
  440. \end_inset
  441. \end_layout
  442. \begin_deeper
  443. \begin_layout Itemize
  444. Completely reimplement analysis from scratch as a reproducible workflow
  445. \end_layout
  446. \begin_layout Itemize
  447. Use newly published methods & algorithms not available during the original
  448. analysis: SICER, csaw, MOFA, ComBat, sva, GREAT, and more
  449. \end_layout
  450. \end_deeper
  451. \begin_layout Itemize
  452. SICER, IDR, csaw, & GREAT to call ChIP-seq peaks genome-wide, perform differenti
  453. al abundance analysis, and relate those peaks to gene expression
  454. \end_layout
  455. \begin_layout Itemize
  456. Promoter counts in sliding windows around each gene's highest-expressed
  457. TSS to investigate coverage distribution within promoters
  458. \end_layout
  459. \begin_layout Section
  460. Results
  461. \end_layout
  462. \begin_layout Standard
  463. \begin_inset Note Note
  464. status open
  465. \begin_layout Plain Layout
  466. Focus on what hypotheses were tested, then select figures that show how
  467. those hypotheses were tested, even if the result is a negative.
  468. \end_layout
  469. \end_inset
  470. \end_layout
  471. \begin_layout Subsection
  472. H3K4 and H3K27 methylation occur in broad regions and are enriched near
  473. promoters
  474. \end_layout
  475. \begin_layout Itemize
  476. Figures comparing MACS (non-broad peak caller) to SICER/epic (broad peak
  477. caller)
  478. \end_layout
  479. \begin_deeper
  480. \begin_layout Itemize
  481. Compare peak sizes and number of called peaks
  482. \end_layout
  483. \begin_layout Itemize
  484. Show representative IDR consistency plots for both
  485. \end_layout
  486. \end_deeper
  487. \begin_layout Itemize
  488. IDR analysis shows that SICER-called peaks are much more reproducible between
  489. biological replicates
  490. \end_layout
  491. \begin_layout Itemize
  492. Each histone mark is enriched within a certain radius of gene TSS positions,
  493. but that radius is different for each mark (figure)
  494. \end_layout
  495. \begin_layout Subsection
  496. RNA-seq has a large confounding batch effect
  497. \end_layout
  498. \begin_layout Itemize
  499. RNA-seq batch effect can be partially corrected, but still induces uncorrectable
  500. biases in downstream analysis
  501. \end_layout
  502. \begin_deeper
  503. \begin_layout Itemize
  504. Figure showing MDS plot before & after ComBat
  505. \end_layout
  506. \begin_layout Itemize
  507. Figure relating sample weights to batches, cell types, time points, etc.,
  508. showing that one batch is significantly worse quality
  509. \end_layout
  510. \begin_layout Itemize
  511. Figures showing p-value histograms for within-batch and cross-batch contrasts,
  512. showing that cross-batch contrasts have attenuated signal, as do comparisons
  513. within the bad batch
  514. \end_layout
  515. \end_deeper
  516. \begin_layout Subsection
  517. ChIP-seq must be corrected for hidden confounding factors
  518. \end_layout
  519. \begin_layout Itemize
  520. Figures showing pre- and post-SVA MDS plots for each histone mark
  521. \end_layout
  522. \begin_layout Itemize
  523. Figures showing BCV plots with and without SVA for each histone mark
  524. \end_layout
  525. \begin_layout Subsection
  526. H3K4 and H3K27 promoter methylation has broadly the expected correlation
  527. with gene expression
  528. \end_layout
  529. \begin_layout Itemize
  530. H3K4 is correlated with higher expression, and H3K27 is correlated with
  531. lower expression genome-wide
  532. \end_layout
  533. \begin_layout Itemize
  534. Figures showing these correlations: box/violin plots of expression distributions
  535. with every combination of peak presence/absence in promoter
  536. \end_layout
  537. \begin_layout Itemize
  538. Appropriate statistical tests showing significant differences in expected
  539. directions
  540. \end_layout
  541. \begin_layout Subsection
  542. MOFA recovers biologically relevant variation from blind analysis by correlating
  543. across datasets
  544. \end_layout
  545. \begin_layout Itemize
  546. MOFA
  547. \begin_inset CommandInset citation
  548. LatexCommand cite
  549. key "Argelaguet2018"
  550. literal "false"
  551. \end_inset
  552. successfully separates biologically relevant patterns of variation from
  553. technical confounding factors without knowing the sample labels, by finding
  554. latent factors that explain variation across multiple data sets.
  555. \end_layout
  556. \begin_deeper
  557. \begin_layout Itemize
  558. Figure: show percent-variance-explained plot from MOFA and PCA-like plots
  559. for the relevant latent factors
  560. \end_layout
  561. \begin_layout Itemize
  562. MOFA analysis also shows that batch effect correction can't get much better
  563. than it already is (Figure comparing blind MOFA batch correction to ComBat
  564. correction)
  565. \end_layout
  566. \end_deeper
  567. \begin_layout Subsection
  568. Naive-to-memory convergence observed in H3K4 and RNA-seq data, not in H3K27me3
  569. \end_layout
  570. \begin_layout Itemize
  571. H3K4 and RNA-seq data show clear evidence of naive convergence with memory
  572. between days 1 and 5 (MDS plot figure, also compare with last figure from
  573. \begin_inset CommandInset citation
  574. LatexCommand cite
  575. key "LaMere2016"
  576. literal "false"
  577. \end_inset
  578. )
  579. \end_layout
  580. \begin_layout Standard
  581. \begin_inset Flex TODO Note (inline)
  582. status open
  583. \begin_layout Plain Layout
  584. Get explicit permission from Sarah to include the figure
  585. \end_layout
  586. \end_inset
  587. \end_layout
  588. \begin_layout Itemize
  589. Table of numbers of genes different between N & M at each time point, showing
  590. dwindling differences at later time points, consistent with convergence
  591. \end_layout
  592. \begin_layout Itemize
  593. Similar figure for H3K27me3 showing lack of convergence
  594. \end_layout
  595. \begin_layout Subsection
  596. Effect of promoter coverage upstream vs downstream of TSS
  597. \end_layout
  598. \begin_layout Itemize
  599. H3K4me peaks seem to correlate with increased expression as long as they
  600. are anywhere near the TSS
  601. \end_layout
  602. \begin_layout Itemize
  603. H3K27me3 peaks can have different correlations to gene expression depending
  604. on their position relative to TSS (e.g.
  605. upstream vs downstream) Results consistent with
  606. \begin_inset CommandInset citation
  607. LatexCommand cite
  608. key "Young2011"
  609. literal "false"
  610. \end_inset
  611. \end_layout
  612. \begin_layout Section
  613. Discussion
  614. \end_layout
  615. \begin_layout Itemize
  616. "Promoter radius" is not constant and must be defined empirically for a
  617. given data set
  618. \end_layout
  619. \begin_layout Itemize
  620. MOFA shows great promise for accelerating discovery of major biological
  621. effects in multi-omics datasets
  622. \end_layout
  623. \begin_deeper
  624. \begin_layout Itemize
  625. MOFA was added to this analysis late and played primarily a confirmatory
  626. role, but it was able to confirm earlier conclusions with much less prior
  627. information (no sample labels) and much less analyst effort
  628. \end_layout
  629. \begin_layout Itemize
  630. MOFA confirmed that the already-implemented batch correction in the RNA-seq
  631. data was already performing as well as possible given the limitations of
  632. the data
  633. \end_layout
  634. \end_deeper
  635. \begin_layout Itemize
  636. Naive-to-memory convergence implies that naive cells are differentiating
  637. into memory cells, and that gene expression and H3K4 methylation are involved
  638. in this differentiation while H3K27me3 is less involved
  639. \end_layout
  640. \begin_layout Itemize
  641. H3K27me3, canonically regarded as a deactivating mark, seems to have a more
  642. complex
  643. \end_layout
  644. \begin_layout Itemize
  645. Discuss advantages of developing using a reproducible workflow
  646. \end_layout
  647. \begin_layout Chapter
  648. Improving array-based analyses of transplant rejection by optimizing data
  649. preprocessing
  650. \end_layout
  651. \begin_layout Standard
  652. \begin_inset Note Note
  653. status open
  654. \begin_layout Plain Layout
  655. Author list: Me, Sunil, Tom, Padma, Dan
  656. \end_layout
  657. \end_inset
  658. \end_layout
  659. \begin_layout Section
  660. Approach
  661. \end_layout
  662. \begin_layout Subsection
  663. Proper pre-processing is essential for array data
  664. \end_layout
  665. \begin_layout Standard
  666. \begin_inset Flex TODO Note (inline)
  667. status open
  668. \begin_layout Plain Layout
  669. This section could probably use some citations
  670. \end_layout
  671. \end_inset
  672. \end_layout
  673. \begin_layout Standard
  674. Microarrays, bead ararys, and similar assays produce raw data in the form
  675. of fluorescence intensity measurements, with the each intensity measurement
  676. proportional to the abundance of some fluorescently-labelled target DNA
  677. or RNA sequence that base pairs to a specific probe sequence.
  678. However, these measurements for each probe are also affected my many technical
  679. confounding factors, such as the concentration of target material, strength
  680. of off-target binding, and the sensitivity of the imaging sensor.
  681. Some array designs also use multiple probe sequences for each target.
  682. Hence, extensive pre-processing of array data is necessary to normalize
  683. out the effects of these technical factors and summarize the information
  684. from multiple probes to arrive at a single usable estimate of abundance
  685. or other relevant quantity, such as a ratio of two abundances, for each
  686. target.
  687. \end_layout
  688. \begin_layout Standard
  689. The choice of pre-processing algorithms used in the analysis of an array
  690. data set can have a large effect on the results of that analysis.
  691. However, despite their importance, these steps are often neglected or rushed
  692. in order to get to the more scientifically interesting analysis steps involving
  693. the actual biology of the system under study.
  694. Hence, it is often possible to achieve substantial gains in statistical
  695. power, model goodness-of-fit, or other relevant performance measures, by
  696. checking the assumptions made by each preprocessing step and choosing specific
  697. normalization methods tailored to the specific goals of the current analysis.
  698. \end_layout
  699. \begin_layout Subsection
  700. Frozen RMA for clinical microarray classifiers
  701. \end_layout
  702. \begin_layout Subsubsection
  703. Standard normalization methods are unsuitable for clinical application
  704. \end_layout
  705. \begin_layout Standard
  706. As the cost of performing microarray assays falls, there is increasing interest
  707. in using genomic assays for diagnostic purposes, such as distinguishing
  708. healthy transplants (TX) from transplants undergoing acute rejection (AR)
  709. or acute dysfunction with no rejection (ADNR).
  710. However, the the standard normalization algorithm used for microarray data,
  711. Robust Multi-chip Average (RMA)
  712. \begin_inset CommandInset citation
  713. LatexCommand cite
  714. key "Irizarry2003a"
  715. literal "false"
  716. \end_inset
  717. , is not applicable in a clinical setting.
  718. Two of the steps in RMA, quantile normalization and probe summarization
  719. by median polish, depend on every array in the data set being normalized.
  720. This means that adding or removing any arrays from a data set changes the
  721. normalized values for all arrays, and data sets that have been normalized
  722. separately cannot be compared to each other.
  723. Hence, when using RMA, any arrays to be analyzed together must also be
  724. normalized together, and the set of arrays included in the data set must
  725. be held constant throughout an analysis.
  726. \end_layout
  727. \begin_layout Standard
  728. These limitations present serious impediments to the use of arrays as a
  729. diagnostic tool.
  730. When training a classifier, the samples to be classified must not be involved
  731. in any step of the training process, lest their inclusion bias the training
  732. process.
  733. Once a classifier is deployed in a clinical setting, the samples to be
  734. classified will not even
  735. \emph on
  736. exist
  737. \emph default
  738. at the time of training, so including them would be impossible even if
  739. it were statistically justifiable.
  740. Therefore, any machine learning application for microarrays demands that
  741. the normalized expression values computed for an array must depend only
  742. on information contained within that array.
  743. This would ensure that each array's normalization is independent of every
  744. other array, and that arrays normalized separately can still be compared
  745. to each other without bias.
  746. \end_layout
  747. \begin_layout Subsubsection
  748. Frozen RMA satisfies clinical normalization requirements
  749. \end_layout
  750. \begin_layout Standard
  751. Frozen RMA (fRMA) addresses these concerns by replacing the quantile normalizati
  752. on and median polish with alternatives that do not introduce inter-array
  753. dependence, allowing each array to be normalized independently of all others
  754. \begin_inset CommandInset citation
  755. LatexCommand cite
  756. key "McCall2010"
  757. literal "false"
  758. \end_inset
  759. .
  760. Quantile normalization is performed against a pre-generated set of quantiles
  761. learned from a collection of 850 publically available arrays sampled from
  762. a wide variety of tissues in the Gene Expression Omnibus (GEO).
  763. Each array's probe intensity distribution is normalized against these pre-gener
  764. ated quantiles.
  765. The median polish step is replaced with a robust weighted average of probe
  766. intensities, using inverse variance weights learned from the same public
  767. GEO data.
  768. The result is a normalization that satisfies the requirements mentioned
  769. above: each array is normalized independently of all others, and any two
  770. normalized arrays can be compared directly to each other.
  771. \end_layout
  772. \begin_layout Standard
  773. One important limitation of fRMA is that it requires a separate reference
  774. data set from which to learn the parameters (reference quantiles and probe
  775. weights) that will be used to normalize each array.
  776. These parameters are specific to a given array platform, and pre-generated
  777. parameters are only provided for the most common platforms, such as Affymetrix
  778. hgu133plus2.
  779. For a less common platform, such as hthgu133pluspm, is is necessary to
  780. learn custom parameters from in-house data before fRMA can be used to normalize
  781. samples on that platform
  782. \begin_inset CommandInset citation
  783. LatexCommand cite
  784. key "HudsonK.&RemediosC.2010"
  785. literal "false"
  786. \end_inset
  787. .
  788. \end_layout
  789. \begin_layout Subsection
  790. Adapting voom to model heteroskedasticity in methylation array data
  791. \end_layout
  792. \begin_layout Subsubsection
  793. Methylation array preprocessing induces heteroskedasticity
  794. \end_layout
  795. \begin_layout Standard
  796. DNA methylation arrays are a relatively new kind of assay that uses microarrays
  797. to measure the degree of methylation on cytosines in specific regions arrayed
  798. across the genome.
  799. First, bisulfite treatment converts all unmethylated cytosines to uracil
  800. (which then become thymine after amplication) while leaving methylated
  801. cytosines unaffected.
  802. Then, each target region is interrogated with two probes: one binds to
  803. the original genomic sequence and interrogates the level of methylated
  804. DNA, and the other binds to the sequence with all Cs replaced by Ts and
  805. interrogates the level of unmethylated DNA.
  806. \end_layout
  807. \begin_layout Standard
  808. \begin_inset Float figure
  809. wide false
  810. sideways false
  811. status collapsed
  812. \begin_layout Plain Layout
  813. \begin_inset Graphics
  814. filename graphics/methylvoom/sigmoid.pdf
  815. \end_inset
  816. \end_layout
  817. \begin_layout Plain Layout
  818. \begin_inset Caption Standard
  819. \begin_layout Plain Layout
  820. \begin_inset CommandInset label
  821. LatexCommand label
  822. name "fig:Sigmoid-beta-m-mapping"
  823. \end_inset
  824. \series bold
  825. Sigmoid shape of the mapping between β and M values
  826. \end_layout
  827. \end_inset
  828. \end_layout
  829. \end_inset
  830. \end_layout
  831. \begin_layout Standard
  832. After normalization, these two probe intensities are summarized in one of
  833. two ways, each with advantages and disadvantages.
  834. β
  835. \series bold
  836. \series default
  837. values, interpreted as fraction of DNA copies methylated, range from 0 to
  838. 1.
  839. β
  840. \series bold
  841. \series default
  842. values are conceptually easy to interpret, but the constrained range makes
  843. them unsuitable for linear modeling, and their error distributions are
  844. highly non-normal, which also frustrates linear modeling.
  845. M-values, interpreted as the log ratio of methylated to unmethylated copies,
  846. are computed by mapping the beta values from
  847. \begin_inset Formula $[0,1]$
  848. \end_inset
  849. onto
  850. \begin_inset Formula $(-\infty,+\infty)$
  851. \end_inset
  852. using a sigmoid curve (Figure
  853. \begin_inset CommandInset ref
  854. LatexCommand ref
  855. reference "fig:Sigmoid-beta-m-mapping"
  856. plural "false"
  857. caps "false"
  858. noprefix "false"
  859. \end_inset
  860. ).
  861. This transformation results in values with better statistical perperties:
  862. the unconstrained range is suitable for linear modeling, and the error
  863. distributions are more normal.
  864. Hence, most linear modeling and other statistical testing on methylation
  865. arrays is performed using M-values.
  866. \end_layout
  867. \begin_layout Standard
  868. However, the steep slope of the sigmoid transformation near 0 and 1 tends
  869. to over-exaggerate small differences in β values near those extremes, which
  870. in turn amplifies the error in those values, leading to a U-shaped trend
  871. in the mean-variance curve: extreme values have higher variances than values
  872. near the middle.
  873. This mean-variance dependency must be accounted for when fitting the linear
  874. model for differential methylation, or else the variance will be systematically
  875. overestimated for probes with moderate M-values and underestimated for
  876. probes with extreme M-values.
  877. \end_layout
  878. \begin_layout Subsubsection
  879. The voom method for RNA-seq data can model M-value heteroskedasticity
  880. \end_layout
  881. \begin_layout Standard
  882. RNA-seq read count data are also known to show heteroskedasticity, and the
  883. voom method was developed for modeling this heteroskedasticity by estimating
  884. the mean-variance trend in the data and using this trend to assign precision
  885. weights to each observation
  886. \begin_inset CommandInset citation
  887. LatexCommand cite
  888. key "Law2013"
  889. literal "false"
  890. \end_inset
  891. .
  892. While methylation array data are not derived from counts and have a very
  893. different mean-variance relationship from that of typical RNA-seq data,
  894. the voom method makes no specific assumptions on the shape of the mean-variance
  895. relationship - it only assumes that the relationship is smooth enough to
  896. model using a lowess curve.
  897. Hence, the method is sufficiently general to model the mean-variance relationsh
  898. ip in methylation array data.
  899. However, the standard implementation of voom assumes that the input is
  900. given in raw read counts, and it must be adapted to run on methylation
  901. M-values.
  902. \end_layout
  903. \begin_layout Standard
  904. \begin_inset Flex TODO Note (inline)
  905. status open
  906. \begin_layout Plain Layout
  907. Put code on Github and reference it
  908. \end_layout
  909. \end_inset
  910. \end_layout
  911. \begin_layout Section
  912. Methods
  913. \end_layout
  914. \begin_layout Subsection
  915. fRMA
  916. \end_layout
  917. \begin_layout Standard
  918. For testing RMA against fRMA, a data set of 157 hgu133plus2 arrays was used,
  919. consisting of blood samples from kidney transplant patients whose grafts
  920. had been graded as TX, AR, or ADNR via biopsy and histology
  921. \begin_inset CommandInset citation
  922. LatexCommand cite
  923. key "Kurian2014"
  924. literal "true"
  925. \end_inset
  926. .
  927. These were split into a training set (23 TX, 35 AR, 21 ADNR) and a validation
  928. set (23 TX, 34 AR, 21 ADNR).
  929. Additionally, an external validation was gathered from public GEO data
  930. (37 TX, 38 AR, no ADNR).
  931. \end_layout
  932. \begin_layout Standard
  933. \begin_inset Flex TODO Note (inline)
  934. status collapsed
  935. \begin_layout Plain Layout
  936. Find appropriate GEO identifiers if possible.
  937. Kurian 2014 says GSE15296, but this seems to be different data.
  938. I also need to look up the GEO accession for the external validation set.
  939. \end_layout
  940. \end_inset
  941. \end_layout
  942. \begin_layout Itemize
  943. Expression array normalization for detecting acute rejection
  944. \end_layout
  945. \begin_layout Itemize
  946. Use frozen RMA, a single-channel variant of RMA
  947. \end_layout
  948. \begin_layout Itemize
  949. Generate custom fRMA normalization vectors for each tissue (biopsy, blood)
  950. \end_layout
  951. \begin_layout Subsubsection
  952. Methylation arrays
  953. \end_layout
  954. \begin_layout Itemize
  955. Methylation arrays for differential methylation in rejection vs.
  956. healthy transplant
  957. \end_layout
  958. \begin_layout Itemize
  959. Adapt voom method originally designed for RNA-seq to model mean-variance
  960. dependence
  961. \end_layout
  962. \begin_layout Itemize
  963. Use sample precision weighting, duplicateCorrelation, and sva to adjust
  964. for other confounding factors
  965. \end_layout
  966. \begin_layout Section
  967. Results
  968. \end_layout
  969. \begin_layout Standard
  970. \begin_inset Flex TODO Note (inline)
  971. status open
  972. \begin_layout Plain Layout
  973. Improve subsection titles in this section
  974. \end_layout
  975. \end_inset
  976. \end_layout
  977. \begin_layout Subsection
  978. fRMA eliminates unwanted dependence of classifier training on normalization
  979. strategy caused by RMA
  980. \end_layout
  981. \begin_layout Subsubsection
  982. Separate normalization with RMA introduces unwanted biases in classification
  983. \end_layout
  984. \begin_layout Standard
  985. \begin_inset Float figure
  986. wide false
  987. sideways false
  988. status collapsed
  989. \begin_layout Plain Layout
  990. \begin_inset Graphics
  991. filename graphics/PAM/predplot.pdf
  992. \end_inset
  993. \end_layout
  994. \begin_layout Plain Layout
  995. \begin_inset Caption Standard
  996. \begin_layout Plain Layout
  997. \begin_inset CommandInset label
  998. LatexCommand label
  999. name "fig:Classifier-probabilities-RMA"
  1000. \end_inset
  1001. \series bold
  1002. Classifier probabilities on validation samples when normalized with RMA
  1003. together vs.
  1004. separately.
  1005. \end_layout
  1006. \end_inset
  1007. \end_layout
  1008. \end_inset
  1009. \end_layout
  1010. \begin_layout Standard
  1011. The initial data set for testing fRMA consisted of 157 hgu133plus2 arrays,
  1012. split into a training set (23 TX, 35 AR, 21 ADNR) and a validation set
  1013. (23 TX, 34 AR, 21 ADNR), along with an external validation set gathered
  1014. from public GEO data (37 TX, 38 AR, no ADNR)
  1015. \begin_inset CommandInset citation
  1016. LatexCommand cite
  1017. key "Kurian2014"
  1018. literal "true"
  1019. \end_inset
  1020. .
  1021. To demonstrate the problem, we considered the problem of training a classifier
  1022. to distinguish TX from AR using the TX and AR samples from the training
  1023. set and validation set as training data, evaluating performance on the
  1024. external validation set.
  1025. First, training and evaluation were performed after normalizing all array
  1026. samples together as a single set using RMA, and second, the internal samples
  1027. were normalized separately from the external samples and the training and
  1028. evaluation were repeated.
  1029. For each sample in the validation set, the classifier probabilities from
  1030. both classifiers were plotted against each other (Fig.
  1031. \begin_inset CommandInset ref
  1032. LatexCommand ref
  1033. reference "fig:Classifier-probabilities-RMA"
  1034. plural "false"
  1035. caps "false"
  1036. noprefix "false"
  1037. \end_inset
  1038. ).
  1039. As expected, separate normalization biases the classifier probabilities,
  1040. resulting in several misclassifications.
  1041. In this case, the bias from separate normalization causes the classifier
  1042. to assign a lower probability of AR to every sample.
  1043. Because it is not feasible to normalize all samples together in a clinical
  1044. context, this shows that an alternative to RMA is required.
  1045. \end_layout
  1046. \begin_layout Subsubsection
  1047. fRMA achieves equal classification performance while eliminating dependence
  1048. on normalization strategy
  1049. \end_layout
  1050. \begin_layout Standard
  1051. \begin_inset Flex TODO Note (inline)
  1052. status open
  1053. \begin_layout Plain Layout
  1054. Cite ROCR: bioinformatics.oxfordjournals.org/cgi/content/abstract/21/20/3940
  1055. \end_layout
  1056. \begin_layout Plain Layout
  1057. Or maybe pROC? https://bmcbioinformatics.biomedcentral.com/articles/10.1186/1471-21
  1058. 05-12-77
  1059. \end_layout
  1060. \end_inset
  1061. \end_layout
  1062. \begin_layout Standard
  1063. \begin_inset Float figure
  1064. wide false
  1065. sideways false
  1066. status open
  1067. \begin_layout Plain Layout
  1068. \begin_inset Graphics
  1069. filename graphics/PAM/external-roc-frma.pdf
  1070. \end_inset
  1071. \end_layout
  1072. \begin_layout Plain Layout
  1073. \begin_inset Caption Standard
  1074. \begin_layout Plain Layout
  1075. \begin_inset CommandInset label
  1076. LatexCommand label
  1077. name "fig:ROC-curve-PAM"
  1078. \end_inset
  1079. ROC curve for PAM on external validation data, normalizing with RMA and
  1080. fRMA
  1081. \end_layout
  1082. \end_inset
  1083. \end_layout
  1084. \end_inset
  1085. \end_layout
  1086. \begin_layout Itemize
  1087. fRMA eliminates this issue by normalizing each sample independently to the
  1088. same quantile distribution and summarizing probes using the same weights.
  1089. \end_layout
  1090. \begin_layout Itemize
  1091. Classifier performance on validation set is identical for
  1092. \begin_inset Quotes eld
  1093. \end_inset
  1094. RMA together
  1095. \begin_inset Quotes erd
  1096. \end_inset
  1097. and fRMA, so switching to clinically applicable normalization does not
  1098. sacrifice accuracy
  1099. \end_layout
  1100. \begin_layout Standard
  1101. \begin_inset Flex TODO Note (inline)
  1102. status open
  1103. \begin_layout Plain Layout
  1104. Check the published paper for any other possibly relevant figures to include
  1105. here.
  1106. \end_layout
  1107. \end_inset
  1108. \end_layout
  1109. \begin_layout Subsection
  1110. fRMA with custom-generated vectors
  1111. \end_layout
  1112. \begin_layout Itemize
  1113. Non-standard platform hthgu133pluspm - no pre-built fRMA vectors available,
  1114. so custom vectors must be learned from in-house data
  1115. \end_layout
  1116. \begin_layout Standard
  1117. \begin_inset Float figure
  1118. wide false
  1119. sideways false
  1120. status collapsed
  1121. \begin_layout Plain Layout
  1122. \begin_inset Graphics
  1123. filename graphics/frma-pax-bx/batchsize_batches.pdf
  1124. \end_inset
  1125. \end_layout
  1126. \begin_layout Plain Layout
  1127. \begin_inset Caption Standard
  1128. \begin_layout Plain Layout
  1129. \begin_inset CommandInset label
  1130. LatexCommand label
  1131. name "fig:batch-size-batches"
  1132. \end_inset
  1133. Effect of batch size selection on number of batches included in fRMA probe
  1134. weight learning
  1135. \end_layout
  1136. \end_inset
  1137. \end_layout
  1138. \end_inset
  1139. \end_layout
  1140. \begin_layout Standard
  1141. \begin_inset Float figure
  1142. wide false
  1143. sideways false
  1144. status collapsed
  1145. \begin_layout Plain Layout
  1146. \begin_inset Graphics
  1147. filename graphics/frma-pax-bx/batchsize_samples.pdf
  1148. \end_inset
  1149. \end_layout
  1150. \begin_layout Plain Layout
  1151. \begin_inset Caption Standard
  1152. \begin_layout Plain Layout
  1153. \begin_inset CommandInset label
  1154. LatexCommand label
  1155. name "fig:batch-size-samples"
  1156. \end_inset
  1157. Effect of batch size selection on number of samples included in fRMA probe
  1158. weight learning
  1159. \end_layout
  1160. \end_inset
  1161. \end_layout
  1162. \end_inset
  1163. \end_layout
  1164. \begin_layout Itemize
  1165. Large body of data available for training fRMA: 341 kidney graft biopsy
  1166. samples, 965 blood samples from graft recipients
  1167. \end_layout
  1168. \begin_deeper
  1169. \begin_layout Itemize
  1170. But not all samples can be used (see trade-off figure)
  1171. \end_layout
  1172. \begin_layout Itemize
  1173. Figure showing trade-off between more samples per group and fewer groups
  1174. with that may samples, to justify choice of number of samples per group
  1175. \end_layout
  1176. \begin_layout Itemize
  1177. pre-generated normalization vectors use ~850 samples
  1178. \begin_inset Flex TODO Note (Margin)
  1179. status collapsed
  1180. \begin_layout Plain Layout
  1181. Look up the exact numbers
  1182. \end_layout
  1183. \end_inset
  1184. \begin_inset CommandInset citation
  1185. LatexCommand cite
  1186. key "McCall2010"
  1187. literal "false"
  1188. \end_inset
  1189. , but are designed to be general across all tissues.
  1190. The samples we have are suitable for tissue-specific normalization vectors.
  1191. \end_layout
  1192. \end_deeper
  1193. \begin_layout Itemize
  1194. Figure: MA plot, RMA vs fRMA, to show that the normalization is appreciably
  1195. and non-linearly different
  1196. \end_layout
  1197. \begin_layout Itemize
  1198. Figure MA plot, fRMA vs fRMA with different randomly-chosen sample subsets
  1199. to show consistency
  1200. \end_layout
  1201. \begin_layout Itemize
  1202. custom fRMA normalization improved cross-validated classifier performance
  1203. \end_layout
  1204. \begin_layout Standard
  1205. \begin_inset Flex TODO Note (inline)
  1206. status open
  1207. \begin_layout Plain Layout
  1208. Get a figure from Tom showing classifier performance improvement (compared
  1209. to all-sample RMA, I guess?), if possible
  1210. \end_layout
  1211. \end_inset
  1212. \end_layout
  1213. \begin_layout Subsection
  1214. Adapting voom to methylation array data improves model fit
  1215. \end_layout
  1216. \begin_layout Itemize
  1217. voom, precision weights, and sva improved model fit
  1218. \end_layout
  1219. \begin_deeper
  1220. \begin_layout Itemize
  1221. Also increased sensitivity for detecting differential methylation
  1222. \end_layout
  1223. \end_deeper
  1224. \begin_layout Itemize
  1225. Figure showing (a) heteroskedasticy without voom, (b) voom-modeled mean-variance
  1226. trend, and (c) homoskedastic mean-variance trend after running voom
  1227. \end_layout
  1228. \begin_layout Itemize
  1229. Figure showing sample weights and their relations to
  1230. \end_layout
  1231. \begin_layout Itemize
  1232. Figure showing MDS plot with and without SVA correction
  1233. \end_layout
  1234. \begin_layout Itemize
  1235. Figure and/or table showing improved p-value historgrams/number of significant
  1236. genes (might need to get this from Padma)
  1237. \end_layout
  1238. \begin_layout Section
  1239. Discussion
  1240. \end_layout
  1241. \begin_layout Itemize
  1242. fRMA enables classifying new samples without re-normalizing the entire data
  1243. set
  1244. \end_layout
  1245. \begin_deeper
  1246. \begin_layout Itemize
  1247. Critical for translating a classifier into clinical practice
  1248. \end_layout
  1249. \end_deeper
  1250. \begin_layout Itemize
  1251. Methods like voom designed for RNA-seq can also help with array analysis
  1252. \end_layout
  1253. \begin_layout Itemize
  1254. Extracting and modeling confounders common to many features improves model
  1255. correspondence to known biology
  1256. \end_layout
  1257. \begin_layout Chapter
  1258. Globin-blocking for more effective blood RNA-seq analysis in primate animal
  1259. model
  1260. \end_layout
  1261. \begin_layout Standard
  1262. \begin_inset Flex TODO Note (inline)
  1263. status open
  1264. \begin_layout Plain Layout
  1265. Choose between above and the paper title: Optimizing yield of deep RNA sequencin
  1266. g for gene expression profiling by globin reduction of peripheral blood
  1267. samples from cynomolgus monkeys (Macaca fascicularis).
  1268. \end_layout
  1269. \end_inset
  1270. \end_layout
  1271. \begin_layout Standard
  1272. \begin_inset Flex TODO Note (inline)
  1273. status open
  1274. \begin_layout Plain Layout
  1275. Chapter author list: https://tex.stackexchange.com/questions/156862/displaying-aut
  1276. hor-for-each-chapter-in-book Every chapter gets an author list, which may
  1277. or may not be part of a citation to a published/preprinted paper.
  1278. \end_layout
  1279. \end_inset
  1280. \end_layout
  1281. \begin_layout Standard
  1282. \begin_inset Flex TODO Note (inline)
  1283. status open
  1284. \begin_layout Plain Layout
  1285. Preprint then cite the paper
  1286. \end_layout
  1287. \end_inset
  1288. \end_layout
  1289. \begin_layout Section*
  1290. Abstract
  1291. \end_layout
  1292. \begin_layout Paragraph
  1293. Background
  1294. \end_layout
  1295. \begin_layout Standard
  1296. Primate blood contains high concentrations of globin messenger RNA.
  1297. Globin reduction is a standard technique used to improve the expression
  1298. results obtained by DNA microarrays on RNA from blood samples.
  1299. However, with whole transcriptome RNA-sequencing (RNA-seq) quickly replacing
  1300. microarrays for many applications, the impact of globin reduction for RNA-seq
  1301. has not been previously studied.
  1302. Moreover, no off-the-shelf kits are available for globin reduction in nonhuman
  1303. primates.
  1304. \end_layout
  1305. \begin_layout Paragraph
  1306. Results
  1307. \end_layout
  1308. \begin_layout Standard
  1309. Here we report a protocol for RNA-seq in primate blood samples that uses
  1310. complimentary oligonucleotides to block reverse transcription of the alpha
  1311. and beta globin genes.
  1312. In test samples from cynomolgus monkeys (Macaca fascicularis), this globin
  1313. blocking protocol approximately doubles the yield of informative (non-globin)
  1314. reads by greatly reducing the fraction of globin reads, while also improving
  1315. the consistency in sequencing depth between samples.
  1316. The increased yield enables detection of about 2000 more genes, significantly
  1317. increases the correlation in measured gene expression levels between samples,
  1318. and increases the sensitivity of differential gene expression tests.
  1319. \end_layout
  1320. \begin_layout Paragraph
  1321. Conclusions
  1322. \end_layout
  1323. \begin_layout Standard
  1324. These results show that globin blocking significantly improves the cost-effectiv
  1325. eness of mRNA sequencing in primate blood samples by doubling the yield
  1326. of useful reads, allowing detection of more genes, and improving the precision
  1327. of gene expression measurements.
  1328. Based on these results, a globin reducing or blocking protocol is recommended
  1329. for all RNA-seq studies of primate blood samples.
  1330. \end_layout
  1331. \begin_layout Section
  1332. Approach
  1333. \end_layout
  1334. \begin_layout Standard
  1335. \begin_inset Note Note
  1336. status open
  1337. \begin_layout Plain Layout
  1338. Consider putting some of this in the Intro chapter
  1339. \end_layout
  1340. \begin_layout Itemize
  1341. Cynomolgus monkeys as a model organism
  1342. \end_layout
  1343. \begin_deeper
  1344. \begin_layout Itemize
  1345. Highly related to humans
  1346. \end_layout
  1347. \begin_layout Itemize
  1348. Small size and short life cycle - good research animal
  1349. \end_layout
  1350. \begin_layout Itemize
  1351. Genomics resources still in development
  1352. \end_layout
  1353. \end_deeper
  1354. \begin_layout Itemize
  1355. Inadequacy of existing blood RNA-seq protocols
  1356. \end_layout
  1357. \begin_deeper
  1358. \begin_layout Itemize
  1359. Existing protocols use a separate globin pulldown step, slowing down processing
  1360. \end_layout
  1361. \end_deeper
  1362. \end_inset
  1363. \end_layout
  1364. \begin_layout Standard
  1365. Increasingly, researchers are turning to high-throughput mRNA sequencing
  1366. technologies (RNA-seq) in preference to expression microarrays for analysis
  1367. of gene expression
  1368. \begin_inset CommandInset citation
  1369. LatexCommand cite
  1370. key "Mutz2012"
  1371. literal "false"
  1372. \end_inset
  1373. .
  1374. The advantages are even greater for study of model organisms with no well-estab
  1375. lished array platforms available, such as the cynomolgus monkey (Macaca
  1376. fascicularis).
  1377. High fractions of globin mRNA are naturally present in mammalian peripheral
  1378. blood samples (up to 70% of total mRNA) and these are known to interfere
  1379. with the results of array-based expression profiling
  1380. \begin_inset CommandInset citation
  1381. LatexCommand cite
  1382. key "Winn2010"
  1383. literal "false"
  1384. \end_inset
  1385. .
  1386. The importance of globin reduction for RNA-seq of blood has only been evaluated
  1387. for a deepSAGE protocol on human samples
  1388. \begin_inset CommandInset citation
  1389. LatexCommand cite
  1390. key "Mastrokolias2012"
  1391. literal "false"
  1392. \end_inset
  1393. .
  1394. In the present report, we evaluated globin reduction using custom blocking
  1395. oligonucleotides for deep RNA-seq of peripheral blood samples from a nonhuman
  1396. primate, cynomolgus monkey, using the Illumina technology platform.
  1397. We demonstrate that globin reduction significantly improves the cost-effectiven
  1398. ess of RNA-seq in blood samples.
  1399. Thus, our protocol offers a significant advantage to any investigator planning
  1400. to use RNA-seq for gene expression profiling of nonhuman primate blood
  1401. samples.
  1402. Our method can be generally applied to any species by designing complementary
  1403. oligonucleotide blocking probes to the globin gene sequences of that species.
  1404. Indeed, any highly expressed but biologically uninformative transcripts
  1405. can also be blocked to further increase sequencing efficiency and value
  1406. \begin_inset CommandInset citation
  1407. LatexCommand cite
  1408. key "Arnaud2016"
  1409. literal "false"
  1410. \end_inset
  1411. .
  1412. \end_layout
  1413. \begin_layout Section
  1414. Methods
  1415. \end_layout
  1416. \begin_layout Subsection*
  1417. Sample collection
  1418. \end_layout
  1419. \begin_layout Standard
  1420. All research reported here was done under IACUC-approved protocols at the
  1421. University of Miami and complied with all applicable federal and state
  1422. regulations and ethical principles for nonhuman primate research.
  1423. Blood draws occurred between 16 April 2012 and 18 June 2015.
  1424. The experimental system involved intrahepatic pancreatic islet transplantation
  1425. into Cynomolgus monkeys with induced diabetes mellitus with or without
  1426. concomitant infusion of mesenchymal stem cells.
  1427. Blood was collected at serial time points before and after transplantation
  1428. into PAXgene Blood RNA tubes (PreAnalytiX/Qiagen, Valencia, CA) at the
  1429. precise volume:volume ratio of 2.5 ml whole blood into 6.9 ml of PAX gene
  1430. additive.
  1431. \end_layout
  1432. \begin_layout Subsection*
  1433. Globin Blocking
  1434. \end_layout
  1435. \begin_layout Standard
  1436. Four oligonucleotides were designed to hybridize to the 3’ end of the transcript
  1437. s for Cynomolgus HBA1, HBA2 and HBB, with two hybridization sites for HBB
  1438. and 2 sites for HBA (the chosen sites were identical in both HBA genes).
  1439. All oligos were purchased from Sigma and were entirely composed of 2’O-Me
  1440. bases with a C3 spacer positioned at the 3’ ends to prevent any polymerase
  1441. mediated primer extension.
  1442. \end_layout
  1443. \begin_layout Quote
  1444. HBA1/2 site 1: GCCCACUCAGACUUUAUUCAAAG-C3spacer
  1445. \end_layout
  1446. \begin_layout Quote
  1447. HBA1/2 site 2: GGUGCAAGGAGGGGAGGAG-C3spacer
  1448. \end_layout
  1449. \begin_layout Quote
  1450. HBB site 1: AAUGAAAAUAAAUGUUUUUUAUUAG-C3spacer
  1451. \end_layout
  1452. \begin_layout Quote
  1453. HBB site 2: CUCAAGGCCCUUCAUAAUAUCCC-C3spacer
  1454. \end_layout
  1455. \begin_layout Subsection*
  1456. RNA-seq Library Preparation
  1457. \end_layout
  1458. \begin_layout Standard
  1459. Sequencing libraries were prepared with 200ng total RNA from each sample.
  1460. Polyadenylated mRNA was selected from 200 ng aliquots of cynomologus blood-deri
  1461. ved total RNA using Ambion Dynabeads Oligo(dT)25 beads (Invitrogen) following
  1462. manufacturer’s recommended protocol.
  1463. PolyA selected RNA was then combined with 8 pmol of HBA1/2 (site 1), 8
  1464. pmol of HBA1/2 (site 2), 12 pmol of HBB (site 1) and 12 pmol of HBB (site
  1465. 2) oligonucleotides.
  1466. In addition, 20 pmol of RT primer containing a portion of the Illumina
  1467. adapter sequence (B-oligo-dTV: GAGTTCCTTGGCACCCGAGAATTCCATTTTTTTTTTTTTTTTTTTV)
  1468. and 4 µL of 5X First Strand buffer (250 mM Tris-HCl pH 8.3, 375 mM KCl,
  1469. 15mM MgCl2) were added in a total volume of 15 µL.
  1470. The RNA was fragmented by heating this cocktail for 3 minutes at 95°C and
  1471. then placed on ice.
  1472. This was followed by the addition of 2 µL 0.1 M DTT, 1 µL RNaseOUT, 1 µL
  1473. 10mM dNTPs 10% biotin-16 aminoallyl-2’- dUTP and 10% biotin-16 aminoallyl-2’-
  1474. dCTP (TriLink Biotech, San Diego, CA), 1 µL Superscript II (200U/ µL, Thermo-Fi
  1475. sher).
  1476. A second “unblocked” library was prepared in the same way for each sample
  1477. but replacing the blocking oligos with an equivalent volume of water.
  1478. The reaction was carried out at 25°C for 15 minutes and 42°C for 40 minutes,
  1479. followed by incubation at 75°C for 10 minutes to inactivate the reverse
  1480. transcriptase.
  1481. \end_layout
  1482. \begin_layout Standard
  1483. The cDNA/RNA hybrid molecules were purified using 1.8X Ampure XP beads (Agencourt
  1484. ) following supplier’s recommended protocol.
  1485. The cDNA/RNA hybrid was eluted in 25 µL of 10 mM Tris-HCl pH 8.0, and then
  1486. bound to 25 µL of M280 Magnetic Streptavidin beads washed per recommended
  1487. protocol (Thermo-Fisher).
  1488. After 30 minutes of binding, beads were washed one time in 100 µL 0.1N NaOH
  1489. to denature and remove the bound RNA, followed by two 100 µL washes with
  1490. 1X TE buffer.
  1491. \end_layout
  1492. \begin_layout Standard
  1493. Subsequent attachment of the 5-prime Illumina A adapter was performed by
  1494. on-bead random primer extension of the following sequence (A-N8 primer:
  1495. TTCAGAGTTCTACAGTCCGACGATCNNNNNNNN).
  1496. Briefly, beads were resuspended in a 20 µL reaction containing 5 µM A-N8
  1497. primer, 40mM Tris-HCl pH 7.5, 20mM MgCl2, 50mM NaCl, 0.325U/µL Sequenase
  1498. 2.0 (Affymetrix, Santa Clara, CA), 0.0025U/µL inorganic pyrophosphatase (Affymetr
  1499. ix) and 300 µM each dNTP.
  1500. Reaction was incubated at 22°C for 30 minutes, then beads were washed 2
  1501. times with 1X TE buffer (200µL).
  1502. \end_layout
  1503. \begin_layout Standard
  1504. The magnetic streptavidin beads were resuspended in 34 µL nuclease-free
  1505. water and added directly to a PCR tube.
  1506. The two Illumina protocol-specified PCR primers were added at 0.53 µM (Illumina
  1507. TruSeq Universal Primer 1 and Illumina TruSeq barcoded PCR primer 2), along
  1508. with 40 µL 2X KAPA HiFi Hotstart ReadyMix (KAPA, Willmington MA) and thermocycl
  1509. ed as follows: starting with 98°C (2 min-hold); 15 cycles of 98°C, 20sec;
  1510. 60°C, 30sec; 72°C, 30sec; and finished with a 72°C (2 min-hold).
  1511. \end_layout
  1512. \begin_layout Standard
  1513. PCR products were purified with 1X Ampure Beads following manufacturer’s
  1514. recommended protocol.
  1515. Libraries were then analyzed using the Agilent TapeStation and quantitation
  1516. of desired size range was performed by “smear analysis”.
  1517. Samples were pooled in equimolar batches of 16 samples.
  1518. Pooled libraries were size selected on 2% agarose gels (E-Gel EX Agarose
  1519. Gels; Thermo-Fisher).
  1520. Products were cut between 250 and 350 bp (corresponding to insert sizes
  1521. of 130 to 230 bps).
  1522. Finished library pools were then sequenced on the Illumina NextSeq500 instrumen
  1523. t with 75 base read lengths.
  1524. \end_layout
  1525. \begin_layout Subsection*
  1526. Read alignment and counting
  1527. \end_layout
  1528. \begin_layout Standard
  1529. Reads were aligned to the cynomolgus genome using STAR
  1530. \begin_inset CommandInset citation
  1531. LatexCommand cite
  1532. key "Dobin2013,Wilson2013"
  1533. literal "false"
  1534. \end_inset
  1535. .
  1536. Counts of uniquely mapped reads were obtained for every gene in each sample
  1537. with the “featureCounts” function from the Rsubread package, using each
  1538. of the three possibilities for the “strandSpecific” option: sense, antisense,
  1539. and unstranded
  1540. \begin_inset CommandInset citation
  1541. LatexCommand cite
  1542. key "Liao2014"
  1543. literal "false"
  1544. \end_inset
  1545. .
  1546. A few artifacts in the cynomolgus genome annotation complicated read counting.
  1547. First, no ortholog is annotated for alpha globin in the cynomolgus genome,
  1548. presumably because the human genome has two alpha globin genes with nearly
  1549. identical sequences, making the orthology relationship ambiguous.
  1550. However, two loci in the cynomolgus genome are as “hemoglobin subunit alpha-lik
  1551. e” (LOC102136192 and LOC102136846).
  1552. LOC102136192 is annotated as a pseudogene while LOC102136846 is annotated
  1553. as protein-coding.
  1554. Our globin reduction protocol was designed to include blocking of these
  1555. two genes.
  1556. Indeed, these two genes have almost the same read counts in each library
  1557. as the properly-annotated HBB gene and much larger counts than any other
  1558. gene in the unblocked libraries, giving confidence that reads derived from
  1559. the real alpha globin are mapping to both genes.
  1560. Thus, reads from both of these loci were counted as alpha globin reads
  1561. in all further analyses.
  1562. The second artifact is a small, uncharacterized non-coding RNA gene (LOC1021365
  1563. 91), which overlaps the HBA-like gene (LOC102136192) on the opposite strand.
  1564. If counting is not performed in stranded mode (or if a non-strand-specific
  1565. sequencing protocol is used), many reads mapping to the globin gene will
  1566. be discarded as ambiguous due to their overlap with this ncRNA gene, resulting
  1567. in significant undercounting of globin reads.
  1568. Therefore, stranded sense counts were used for all further analysis in
  1569. the present study to insure that we accurately accounted for globin transcript
  1570. reduction.
  1571. However, we note that stranded reads are not necessary for RNA-seq using
  1572. our protocol in standard practice.
  1573. \end_layout
  1574. \begin_layout Subsection*
  1575. Normalization and Exploratory Data Analysis
  1576. \end_layout
  1577. \begin_layout Standard
  1578. Libraries were normalized by computing scaling factors using the edgeR package’s
  1579. Trimmed Mean of M-values method
  1580. \begin_inset CommandInset citation
  1581. LatexCommand cite
  1582. key "Robinson2010"
  1583. literal "false"
  1584. \end_inset
  1585. .
  1586. Log2 counts per million values (logCPM) were calculated using the cpm function
  1587. in edgeR for individual samples and aveLogCPM function for averages across
  1588. groups of samples, using those functions’ default prior count values to
  1589. avoid taking the logarithm of 0.
  1590. Genes were considered “present” if their average normalized logCPM values
  1591. across all libraries were at least -1.
  1592. Normalizing for gene length was unnecessary because the sequencing protocol
  1593. is 3’-biased and hence the expected read count for each gene is related
  1594. to the transcript’s copy number but not its length.
  1595. \end_layout
  1596. \begin_layout Standard
  1597. In order to assess the effect of blocking on reproducibility, Pearson and
  1598. Spearman correlation coefficients were computed between the logCPM values
  1599. for every pair of libraries within the globin-blocked (GB) and unblocked
  1600. (non-GB) groups, and edgeR's “estimateDisp” function was used to compute
  1601. negative binomial dispersions separately for the two groups
  1602. \begin_inset CommandInset citation
  1603. LatexCommand cite
  1604. key "Chen2014"
  1605. literal "false"
  1606. \end_inset
  1607. .
  1608. \end_layout
  1609. \begin_layout Subsection*
  1610. Differential Expression Analysis
  1611. \end_layout
  1612. \begin_layout Standard
  1613. All tests for differential gene expression were performed using edgeR, by
  1614. first fitting a negative binomial generalized linear model to the counts
  1615. and normalization factors and then performing a quasi-likelihood F-test
  1616. with robust estimation of outlier gene dispersions
  1617. \begin_inset CommandInset citation
  1618. LatexCommand cite
  1619. key "Lund2012,Phipson2016"
  1620. literal "false"
  1621. \end_inset
  1622. .
  1623. To investigate the effects of globin blocking on each gene, an additive
  1624. model was fit to the full data with coefficients for globin blocking and
  1625. SampleID.
  1626. To test the effect of globin blocking on detection of differentially expressed
  1627. genes, the GB samples and non-GB samples were each analyzed independently
  1628. as follows: for each animal with both a pre-transplant and a post-transplant
  1629. time point in the data set, the pre-transplant sample and the earliest
  1630. post-transplant sample were selected, and all others were excluded, yielding
  1631. a pre-/post-transplant pair of samples for each animal (N=7 animals with
  1632. paired samples).
  1633. These samples were analyzed for pre-transplant vs.
  1634. post-transplant differential gene expression while controlling for inter-animal
  1635. variation using an additive model with coefficients for transplant and
  1636. animal ID.
  1637. In all analyses, p-values were adjusted using the Benjamini-Hochberg procedure
  1638. for FDR correction
  1639. \begin_inset CommandInset citation
  1640. LatexCommand cite
  1641. key "Benjamini1995"
  1642. literal "false"
  1643. \end_inset
  1644. .
  1645. \end_layout
  1646. \begin_layout Standard
  1647. \begin_inset Note Note
  1648. status open
  1649. \begin_layout Itemize
  1650. New blood RNA-seq protocol to block reverse transcription of globin genes
  1651. \end_layout
  1652. \begin_layout Itemize
  1653. Blood RNA-seq time course after transplants with/without MSC infusion
  1654. \end_layout
  1655. \end_inset
  1656. \end_layout
  1657. \begin_layout Section
  1658. Results
  1659. \end_layout
  1660. \begin_layout Subsection*
  1661. Globin blocking yields a larger and more consistent fraction of useful reads
  1662. \end_layout
  1663. \begin_layout Standard
  1664. The objective of the present study was to validate a new protocol for deep
  1665. RNA-seq of whole blood drawn into PaxGene tubes from cynomolgus monkeys
  1666. undergoing islet transplantation, with particular focus on minimizing the
  1667. loss of useful sequencing space to uninformative globin reads.
  1668. The details of the analysis with respect to transplant outcomes and the
  1669. impact of mesenchymal stem cell treatment will be reported in a separate
  1670. manuscript (in preparation).
  1671. To focus on the efficacy of our globin blocking protocol, 37 blood samples,
  1672. 16 from pre-transplant and 21 from post-transplant time points, were each
  1673. prepped once with and once without globin blocking oligos, and were then
  1674. sequenced on an Illumina NextSeq500 instrument.
  1675. The number of reads aligning to each gene in the cynomolgus genome was
  1676. counted.
  1677. Table 1 summarizes the distribution of read fractions among the GB and
  1678. non-GB libraries.
  1679. In the libraries with no globin blocking, globin reads made up an average
  1680. of 44.6% of total input reads, while reads assigned to all other genes made
  1681. up an average of 26.3%.
  1682. The remaining reads either aligned to intergenic regions (that include
  1683. long non-coding RNAs) or did not align with any annotated transcripts in
  1684. the current build of the cynomolgus genome.
  1685. In the GB libraries, globin reads made up only 3.48% and reads assigned
  1686. to all other genes increased to 50.4%.
  1687. Thus, globin blocking resulted in a 92.2% reduction in globin reads and
  1688. a 91.6% increase in yield of useful non-globin reads.
  1689. \end_layout
  1690. \begin_layout Standard
  1691. This reduction is not quite as efficient as the previous analysis showed
  1692. for human samples by DeepSAGE (<0.4% globin reads after globin reduction)
  1693. \begin_inset CommandInset citation
  1694. LatexCommand cite
  1695. key "Mastrokolias2012"
  1696. literal "false"
  1697. \end_inset
  1698. .
  1699. Nonetheless, this degree of globin reduction is sufficient to nearly double
  1700. the yield of useful reads.
  1701. Thus, globin blocking cuts the required sequencing effort (and costs) to
  1702. achieve a target coverage depth by almost 50%.
  1703. Consistent with this near doubling of yield, the average difference in
  1704. un-normalized logCPM across all genes between the GB libraries and non-GB
  1705. libraries is approximately 1 (mean = 1.01, median = 1.08), an overall 2-fold
  1706. increase.
  1707. Un-normalized values are used here because the TMM normalization correctly
  1708. identifies this 2-fold difference as biologically irrelevant and removes
  1709. it.
  1710. \end_layout
  1711. \begin_layout Standard
  1712. \begin_inset Float figure
  1713. wide false
  1714. sideways false
  1715. status open
  1716. \begin_layout Plain Layout
  1717. \align center
  1718. \begin_inset Graphics
  1719. filename graphics/Globin Paper/figure1 - globin-fractions.pdf
  1720. \end_inset
  1721. \end_layout
  1722. \begin_layout Plain Layout
  1723. \begin_inset Caption Standard
  1724. \begin_layout Plain Layout
  1725. \series bold
  1726. \begin_inset Argument 1
  1727. status collapsed
  1728. \begin_layout Plain Layout
  1729. Fraction of genic reads in each sample aligned to non-globin genes, with
  1730. and without globin blocking (GB).
  1731. \end_layout
  1732. \end_inset
  1733. \begin_inset CommandInset label
  1734. LatexCommand label
  1735. name "fig:Fraction-of-genic-reads"
  1736. \end_inset
  1737. Fraction of genic reads in each sample aligned to non-globin genes, with
  1738. and without globin blocking (GB).
  1739. \series default
  1740. All reads in each sequencing library were aligned to the cyno genome, and
  1741. the number of reads uniquely aligning to each gene was counted.
  1742. For each sample, counts were summed separately for all globin genes and
  1743. for the remainder of the genes (non-globin genes), and the fraction of
  1744. genic reads aligned to non-globin genes was computed.
  1745. Each point represents an individual sample.
  1746. Gray + signs indicate the means for globin-blocked libraries and unblocked
  1747. libraries.
  1748. The overall distribution for each group is represented as a notched box
  1749. plots.
  1750. Points are randomly spread vertically to avoid excessive overlapping.
  1751. \end_layout
  1752. \end_inset
  1753. \end_layout
  1754. \begin_layout Plain Layout
  1755. \end_layout
  1756. \end_inset
  1757. \end_layout
  1758. \begin_layout Standard
  1759. \begin_inset Float table
  1760. placement p
  1761. wide false
  1762. sideways true
  1763. status open
  1764. \begin_layout Plain Layout
  1765. \align center
  1766. \begin_inset Tabular
  1767. <lyxtabular version="3" rows="4" columns="7">
  1768. <features tabularvalignment="middle">
  1769. <column alignment="center" valignment="top">
  1770. <column alignment="center" valignment="top">
  1771. <column alignment="center" valignment="top">
  1772. <column alignment="center" valignment="top">
  1773. <column alignment="center" valignment="top">
  1774. <column alignment="center" valignment="top">
  1775. <column alignment="center" valignment="top">
  1776. <row>
  1777. <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  1778. \begin_inset Text
  1779. \begin_layout Plain Layout
  1780. \end_layout
  1781. \end_inset
  1782. </cell>
  1783. <cell multicolumn="1" alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  1784. \begin_inset Text
  1785. \begin_layout Plain Layout
  1786. \family roman
  1787. \series medium
  1788. \shape up
  1789. \size normal
  1790. \emph off
  1791. \bar no
  1792. \strikeout off
  1793. \xout off
  1794. \uuline off
  1795. \uwave off
  1796. \noun off
  1797. \color none
  1798. Percent of Total Reads
  1799. \end_layout
  1800. \end_inset
  1801. </cell>
  1802. <cell multicolumn="2" alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  1803. \begin_inset Text
  1804. \begin_layout Plain Layout
  1805. \end_layout
  1806. \end_inset
  1807. </cell>
  1808. <cell multicolumn="2" alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  1809. \begin_inset Text
  1810. \begin_layout Plain Layout
  1811. \end_layout
  1812. \end_inset
  1813. </cell>
  1814. <cell multicolumn="2" alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  1815. \begin_inset Text
  1816. \begin_layout Plain Layout
  1817. \end_layout
  1818. \end_inset
  1819. </cell>
  1820. <cell multicolumn="1" alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
  1821. \begin_inset Text
  1822. \begin_layout Plain Layout
  1823. \family roman
  1824. \series medium
  1825. \shape up
  1826. \size normal
  1827. \emph off
  1828. \bar no
  1829. \strikeout off
  1830. \xout off
  1831. \uuline off
  1832. \uwave off
  1833. \noun off
  1834. \color none
  1835. Percent of Genic Reads
  1836. \end_layout
  1837. \end_inset
  1838. </cell>
  1839. <cell multicolumn="2" alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
  1840. \begin_inset Text
  1841. \begin_layout Plain Layout
  1842. \end_layout
  1843. \end_inset
  1844. </cell>
  1845. </row>
  1846. <row>
  1847. <cell alignment="center" valignment="top" bottomline="true" leftline="true" usebox="none">
  1848. \begin_inset Text
  1849. \begin_layout Plain Layout
  1850. GB
  1851. \end_layout
  1852. \end_inset
  1853. </cell>
  1854. <cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
  1855. \begin_inset Text
  1856. \begin_layout Plain Layout
  1857. \family roman
  1858. \series medium
  1859. \shape up
  1860. \size normal
  1861. \emph off
  1862. \bar no
  1863. \strikeout off
  1864. \xout off
  1865. \uuline off
  1866. \uwave off
  1867. \noun off
  1868. \color none
  1869. Non-globin Reads
  1870. \end_layout
  1871. \end_inset
  1872. </cell>
  1873. <cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
  1874. \begin_inset Text
  1875. \begin_layout Plain Layout
  1876. \family roman
  1877. \series medium
  1878. \shape up
  1879. \size normal
  1880. \emph off
  1881. \bar no
  1882. \strikeout off
  1883. \xout off
  1884. \uuline off
  1885. \uwave off
  1886. \noun off
  1887. \color none
  1888. Globin Reads
  1889. \end_layout
  1890. \end_inset
  1891. </cell>
  1892. <cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
  1893. \begin_inset Text
  1894. \begin_layout Plain Layout
  1895. \family roman
  1896. \series medium
  1897. \shape up
  1898. \size normal
  1899. \emph off
  1900. \bar no
  1901. \strikeout off
  1902. \xout off
  1903. \uuline off
  1904. \uwave off
  1905. \noun off
  1906. \color none
  1907. All Genic Reads
  1908. \end_layout
  1909. \end_inset
  1910. </cell>
  1911. <cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
  1912. \begin_inset Text
  1913. \begin_layout Plain Layout
  1914. \family roman
  1915. \series medium
  1916. \shape up
  1917. \size normal
  1918. \emph off
  1919. \bar no
  1920. \strikeout off
  1921. \xout off
  1922. \uuline off
  1923. \uwave off
  1924. \noun off
  1925. \color none
  1926. All Aligned Reads
  1927. \end_layout
  1928. \end_inset
  1929. </cell>
  1930. <cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
  1931. \begin_inset Text
  1932. \begin_layout Plain Layout
  1933. \family roman
  1934. \series medium
  1935. \shape up
  1936. \size normal
  1937. \emph off
  1938. \bar no
  1939. \strikeout off
  1940. \xout off
  1941. \uuline off
  1942. \uwave off
  1943. \noun off
  1944. \color none
  1945. Non-globin Reads
  1946. \end_layout
  1947. \end_inset
  1948. </cell>
  1949. <cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" rightline="true" usebox="none">
  1950. \begin_inset Text
  1951. \begin_layout Plain Layout
  1952. \family roman
  1953. \series medium
  1954. \shape up
  1955. \size normal
  1956. \emph off
  1957. \bar no
  1958. \strikeout off
  1959. \xout off
  1960. \uuline off
  1961. \uwave off
  1962. \noun off
  1963. \color none
  1964. Globin Reads
  1965. \end_layout
  1966. \end_inset
  1967. </cell>
  1968. </row>
  1969. <row>
  1970. <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  1971. \begin_inset Text
  1972. \begin_layout Plain Layout
  1973. \family roman
  1974. \series medium
  1975. \shape up
  1976. \size normal
  1977. \emph off
  1978. \bar no
  1979. \strikeout off
  1980. \xout off
  1981. \uuline off
  1982. \uwave off
  1983. \noun off
  1984. \color none
  1985. Yes
  1986. \end_layout
  1987. \end_inset
  1988. </cell>
  1989. <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  1990. \begin_inset Text
  1991. \begin_layout Plain Layout
  1992. \family roman
  1993. \series medium
  1994. \shape up
  1995. \size normal
  1996. \emph off
  1997. \bar no
  1998. \strikeout off
  1999. \xout off
  2000. \uuline off
  2001. \uwave off
  2002. \noun off
  2003. \color none
  2004. 50.4% ± 6.82
  2005. \end_layout
  2006. \end_inset
  2007. </cell>
  2008. <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  2009. \begin_inset Text
  2010. \begin_layout Plain Layout
  2011. \family roman
  2012. \series medium
  2013. \shape up
  2014. \size normal
  2015. \emph off
  2016. \bar no
  2017. \strikeout off
  2018. \xout off
  2019. \uuline off
  2020. \uwave off
  2021. \noun off
  2022. \color none
  2023. 3.48% ± 2.94
  2024. \end_layout
  2025. \end_inset
  2026. </cell>
  2027. <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  2028. \begin_inset Text
  2029. \begin_layout Plain Layout
  2030. \family roman
  2031. \series medium
  2032. \shape up
  2033. \size normal
  2034. \emph off
  2035. \bar no
  2036. \strikeout off
  2037. \xout off
  2038. \uuline off
  2039. \uwave off
  2040. \noun off
  2041. \color none
  2042. 53.9% ± 6.81
  2043. \end_layout
  2044. \end_inset
  2045. </cell>
  2046. <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  2047. \begin_inset Text
  2048. \begin_layout Plain Layout
  2049. \family roman
  2050. \series medium
  2051. \shape up
  2052. \size normal
  2053. \emph off
  2054. \bar no
  2055. \strikeout off
  2056. \xout off
  2057. \uuline off
  2058. \uwave off
  2059. \noun off
  2060. \color none
  2061. 89.7% ± 2.40
  2062. \end_layout
  2063. \end_inset
  2064. </cell>
  2065. <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  2066. \begin_inset Text
  2067. \begin_layout Plain Layout
  2068. \family roman
  2069. \series medium
  2070. \shape up
  2071. \size normal
  2072. \emph off
  2073. \bar no
  2074. \strikeout off
  2075. \xout off
  2076. \uuline off
  2077. \uwave off
  2078. \noun off
  2079. \color none
  2080. 93.5% ± 5.25
  2081. \end_layout
  2082. \end_inset
  2083. </cell>
  2084. <cell alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
  2085. \begin_inset Text
  2086. \begin_layout Plain Layout
  2087. \family roman
  2088. \series medium
  2089. \shape up
  2090. \size normal
  2091. \emph off
  2092. \bar no
  2093. \strikeout off
  2094. \xout off
  2095. \uuline off
  2096. \uwave off
  2097. \noun off
  2098. \color none
  2099. 6.49% ± 5.25
  2100. \end_layout
  2101. \end_inset
  2102. </cell>
  2103. </row>
  2104. <row>
  2105. <cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
  2106. \begin_inset Text
  2107. \begin_layout Plain Layout
  2108. \family roman
  2109. \series medium
  2110. \shape up
  2111. \size normal
  2112. \emph off
  2113. \bar no
  2114. \strikeout off
  2115. \xout off
  2116. \uuline off
  2117. \uwave off
  2118. \noun off
  2119. \color none
  2120. No
  2121. \end_layout
  2122. \end_inset
  2123. </cell>
  2124. <cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
  2125. \begin_inset Text
  2126. \begin_layout Plain Layout
  2127. \family roman
  2128. \series medium
  2129. \shape up
  2130. \size normal
  2131. \emph off
  2132. \bar no
  2133. \strikeout off
  2134. \xout off
  2135. \uuline off
  2136. \uwave off
  2137. \noun off
  2138. \color none
  2139. 26.3% ± 8.95
  2140. \end_layout
  2141. \end_inset
  2142. </cell>
  2143. <cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
  2144. \begin_inset Text
  2145. \begin_layout Plain Layout
  2146. \family roman
  2147. \series medium
  2148. \shape up
  2149. \size normal
  2150. \emph off
  2151. \bar no
  2152. \strikeout off
  2153. \xout off
  2154. \uuline off
  2155. \uwave off
  2156. \noun off
  2157. \color none
  2158. 44.6% ± 16.6
  2159. \end_layout
  2160. \end_inset
  2161. </cell>
  2162. <cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
  2163. \begin_inset Text
  2164. \begin_layout Plain Layout
  2165. \family roman
  2166. \series medium
  2167. \shape up
  2168. \size normal
  2169. \emph off
  2170. \bar no
  2171. \strikeout off
  2172. \xout off
  2173. \uuline off
  2174. \uwave off
  2175. \noun off
  2176. \color none
  2177. 70.1% ± 9.38
  2178. \end_layout
  2179. \end_inset
  2180. </cell>
  2181. <cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
  2182. \begin_inset Text
  2183. \begin_layout Plain Layout
  2184. \family roman
  2185. \series medium
  2186. \shape up
  2187. \size normal
  2188. \emph off
  2189. \bar no
  2190. \strikeout off
  2191. \xout off
  2192. \uuline off
  2193. \uwave off
  2194. \noun off
  2195. \color none
  2196. 90.7% ± 5.16
  2197. \end_layout
  2198. \end_inset
  2199. </cell>
  2200. <cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
  2201. \begin_inset Text
  2202. \begin_layout Plain Layout
  2203. \family roman
  2204. \series medium
  2205. \shape up
  2206. \size normal
  2207. \emph off
  2208. \bar no
  2209. \strikeout off
  2210. \xout off
  2211. \uuline off
  2212. \uwave off
  2213. \noun off
  2214. \color none
  2215. 38.8% ± 17.1
  2216. \end_layout
  2217. \end_inset
  2218. </cell>
  2219. <cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" rightline="true" usebox="none">
  2220. \begin_inset Text
  2221. \begin_layout Plain Layout
  2222. \family roman
  2223. \series medium
  2224. \shape up
  2225. \size normal
  2226. \emph off
  2227. \bar no
  2228. \strikeout off
  2229. \xout off
  2230. \uuline off
  2231. \uwave off
  2232. \noun off
  2233. \color none
  2234. 61.2% ± 17.1
  2235. \end_layout
  2236. \end_inset
  2237. </cell>
  2238. </row>
  2239. </lyxtabular>
  2240. \end_inset
  2241. \end_layout
  2242. \begin_layout Plain Layout
  2243. \begin_inset Caption Standard
  2244. \begin_layout Plain Layout
  2245. \series bold
  2246. \begin_inset Argument 1
  2247. status collapsed
  2248. \begin_layout Plain Layout
  2249. Fractions of reads mapping to genomic features in GB and non-GB samples.
  2250. \end_layout
  2251. \end_inset
  2252. \begin_inset CommandInset label
  2253. LatexCommand label
  2254. name "tab:Fractions-of-reads"
  2255. \end_inset
  2256. Fractions of reads mapping to genomic features in GB and non-GB samples.
  2257. \series default
  2258. All values are given as mean ± standard deviation.
  2259. \end_layout
  2260. \end_inset
  2261. \end_layout
  2262. \begin_layout Plain Layout
  2263. \end_layout
  2264. \end_inset
  2265. \end_layout
  2266. \begin_layout Standard
  2267. Another important aspect is that the standard deviations in Table
  2268. \begin_inset CommandInset ref
  2269. LatexCommand ref
  2270. reference "tab:Fractions-of-reads"
  2271. plural "false"
  2272. caps "false"
  2273. noprefix "false"
  2274. \end_inset
  2275. are uniformly smaller in the GB samples than the non-GB ones, indicating
  2276. much greater consistency of yield.
  2277. This is best seen in the percentage of non-globin reads as a fraction of
  2278. total reads aligned to annotated genes (genic reads).
  2279. For the non-GB samples, this measure ranges from 10.9% to 80.9%, while for
  2280. the GB samples it ranges from 81.9% to 99.9% (Figure
  2281. \begin_inset CommandInset ref
  2282. LatexCommand ref
  2283. reference "fig:Fraction-of-genic-reads"
  2284. plural "false"
  2285. caps "false"
  2286. noprefix "false"
  2287. \end_inset
  2288. ).
  2289. This means that for applications where it is critical that each sample
  2290. achieve a specified minimum coverage in order to provide useful information,
  2291. it would be necessary to budget up to 10 times the sequencing depth per
  2292. sample without globin blocking, even though the average yield improvement
  2293. for globin blocking is only 2-fold, because every sample has a chance of
  2294. being 90% globin and 10% useful reads.
  2295. Hence, the more consistent behavior of GB samples makes planning an experiment
  2296. easier and more efficient because it eliminates the need to over-sequence
  2297. every sample in order to guard against the worst case of a high-globin
  2298. fraction.
  2299. \end_layout
  2300. \begin_layout Subsection*
  2301. Globin blocking lowers the noise floor and allows detection of about 2000
  2302. more genes
  2303. \end_layout
  2304. \begin_layout Standard
  2305. \begin_inset Flex TODO Note (inline)
  2306. status open
  2307. \begin_layout Plain Layout
  2308. Remove redundant titles from figures
  2309. \end_layout
  2310. \end_inset
  2311. \end_layout
  2312. \begin_layout Standard
  2313. \begin_inset Float figure
  2314. wide false
  2315. sideways false
  2316. status open
  2317. \begin_layout Plain Layout
  2318. \align center
  2319. \begin_inset Graphics
  2320. filename graphics/Globin Paper/figure2 - aveLogCPM-colored.pdf
  2321. \end_inset
  2322. \end_layout
  2323. \begin_layout Plain Layout
  2324. \begin_inset Caption Standard
  2325. \begin_layout Plain Layout
  2326. \series bold
  2327. \begin_inset Argument 1
  2328. status collapsed
  2329. \begin_layout Plain Layout
  2330. Distributions of average group gene abundances when normalized separately
  2331. or together.
  2332. \end_layout
  2333. \end_inset
  2334. \begin_inset CommandInset label
  2335. LatexCommand label
  2336. name "fig:logcpm-dists"
  2337. \end_inset
  2338. Distributions of average group gene abundances when normalized separately
  2339. or together.
  2340. \series default
  2341. All reads in each sequencing library were aligned to the cyno genome, and
  2342. the number of reads uniquely aligning to each gene was counted.
  2343. Genes with zero counts in all libraries were discarded.
  2344. Libraries were normalized using the TMM method.
  2345. Libraries were split into globin-blocked (GB) and non-GB groups and the
  2346. average abundance for each gene in both groups, measured in log2 counts
  2347. per million reads counted, was computed using the aveLogCPM function.
  2348. The distribution of average gene logCPM values was plotted for both groups
  2349. using a kernel density plot to approximate a continuous distribution.
  2350. The logCPM GB distributions are marked in red, non-GB in blue.
  2351. The black vertical line denotes the chosen detection threshold of -1.
  2352. Top panel: Libraries were split into GB and non-GB groups first and normalized
  2353. separately.
  2354. Bottom panel: Libraries were all normalized together first and then split
  2355. into groups.
  2356. \end_layout
  2357. \end_inset
  2358. \end_layout
  2359. \begin_layout Plain Layout
  2360. \end_layout
  2361. \end_inset
  2362. \end_layout
  2363. \begin_layout Standard
  2364. Since globin blocking yields more usable sequencing depth, it should also
  2365. allow detection of more genes at any given threshold.
  2366. When we looked at the distribution of average normalized logCPM values
  2367. across all libraries for genes with at least one read assigned to them,
  2368. we observed the expected bimodal distribution, with a high-abundance "signal"
  2369. peak representing detected genes and a low-abundance "noise" peak representing
  2370. genes whose read count did not rise above the noise floor (Figure
  2371. \begin_inset CommandInset ref
  2372. LatexCommand ref
  2373. reference "fig:logcpm-dists"
  2374. plural "false"
  2375. caps "false"
  2376. noprefix "false"
  2377. \end_inset
  2378. ).
  2379. Consistent with the 2-fold increase in raw counts assigned to non-globin
  2380. genes, the signal peak for GB samples is shifted to the right relative
  2381. to the non-GB signal peak.
  2382. When all the samples are normalized together, this difference is normalized
  2383. out, lining up the signal peaks, and this reveals that, as expected, the
  2384. noise floor for the GB samples is about 2-fold lower.
  2385. This greater separation between signal and noise peaks in the GB samples
  2386. means that low-expression genes should be more easily detected and more
  2387. precisely quantified than in the non-GB samples.
  2388. \end_layout
  2389. \begin_layout Standard
  2390. \begin_inset Float figure
  2391. wide false
  2392. sideways false
  2393. status open
  2394. \begin_layout Plain Layout
  2395. \align center
  2396. \begin_inset Graphics
  2397. filename graphics/Globin Paper/figure3 - detection.pdf
  2398. \end_inset
  2399. \end_layout
  2400. \begin_layout Plain Layout
  2401. \begin_inset Caption Standard
  2402. \begin_layout Plain Layout
  2403. \series bold
  2404. \begin_inset Argument 1
  2405. status collapsed
  2406. \begin_layout Plain Layout
  2407. Gene detections as a function of abundance thresholds in globin-blocked
  2408. (GB) and non-GB samples.
  2409. \end_layout
  2410. \end_inset
  2411. \begin_inset CommandInset label
  2412. LatexCommand label
  2413. name "fig:Gene-detections"
  2414. \end_inset
  2415. Gene detections as a function of abundance thresholds in globin-blocked
  2416. (GB) and non-GB samples.
  2417. \series default
  2418. Average abundance (logCPM,
  2419. \begin_inset Formula $\log_{2}$
  2420. \end_inset
  2421. counts per million reads counted) was computed by separate group normalization
  2422. as described in Figure
  2423. \begin_inset CommandInset ref
  2424. LatexCommand ref
  2425. reference "fig:logcpm-dists"
  2426. plural "false"
  2427. caps "false"
  2428. noprefix "false"
  2429. \end_inset
  2430. for both the GB and non-GB groups, as well as for all samples considered
  2431. as one large group.
  2432. For each every integer threshold from -2 to 3, the number of genes detected
  2433. at or above that logCPM threshold was plotted for each group.
  2434. \end_layout
  2435. \end_inset
  2436. \end_layout
  2437. \begin_layout Plain Layout
  2438. \end_layout
  2439. \end_inset
  2440. \end_layout
  2441. \begin_layout Standard
  2442. Based on these distributions, we selected a detection threshold of -1, which
  2443. is approximately the leftmost edge of the trough between the signal and
  2444. noise peaks.
  2445. This represents the most liberal possible detection threshold that doesn't
  2446. call substantial numbers of noise genes as detected.
  2447. Among the full dataset, 13429 genes were detected at this threshold, and
  2448. 22276 were not.
  2449. When considering the GB libraries and non-GB libraries separately and re-comput
  2450. ing normalization factors independently within each group, 14535 genes were
  2451. detected in the GB libraries while only 12460 were detected in the non-GB
  2452. libraries.
  2453. Thus, GB allowed the detection of 2000 extra genes that were buried under
  2454. the noise floor without GB.
  2455. This pattern of at least 2000 additional genes detected with GB was also
  2456. consistent across a wide range of possible detection thresholds, from -2
  2457. to 3 (see Figure
  2458. \begin_inset CommandInset ref
  2459. LatexCommand ref
  2460. reference "fig:Gene-detections"
  2461. plural "false"
  2462. caps "false"
  2463. noprefix "false"
  2464. \end_inset
  2465. ).
  2466. \end_layout
  2467. \begin_layout Subsection*
  2468. Globin blocking does not add significant additional noise or decrease sample
  2469. quality
  2470. \end_layout
  2471. \begin_layout Standard
  2472. One potential worry is that the globin blocking protocol could perturb the
  2473. levels of non-globin genes.
  2474. There are two kinds of possible perturbations: systematic and random.
  2475. The former is not a major concern for detection of differential expression,
  2476. since a 2-fold change in every sample has no effect on the relative fold
  2477. change between samples.
  2478. In contrast, random perturbations would increase the noise and obscure
  2479. the signal in the dataset, reducing the capacity to detect differential
  2480. expression.
  2481. \end_layout
  2482. \begin_layout Standard
  2483. \begin_inset Float figure
  2484. wide false
  2485. sideways false
  2486. status open
  2487. \begin_layout Plain Layout
  2488. \align center
  2489. \begin_inset Graphics
  2490. filename graphics/Globin Paper/figure4 - maplot-colored.pdf
  2491. \end_inset
  2492. \end_layout
  2493. \begin_layout Plain Layout
  2494. \begin_inset Caption Standard
  2495. \begin_layout Plain Layout
  2496. \begin_inset Argument 1
  2497. status collapsed
  2498. \begin_layout Plain Layout
  2499. MA plot showing effects of globin blocking on each gene's abundance.
  2500. \end_layout
  2501. \end_inset
  2502. \begin_inset CommandInset label
  2503. LatexCommand label
  2504. name "fig:MA-plot"
  2505. \end_inset
  2506. \series bold
  2507. MA plot showing effects of globin blocking on each gene's abundance.
  2508. \series default
  2509. All libraries were normalized together as described in Figure
  2510. \begin_inset CommandInset ref
  2511. LatexCommand ref
  2512. reference "fig:logcpm-dists"
  2513. plural "false"
  2514. caps "false"
  2515. noprefix "false"
  2516. \end_inset
  2517. , and genes with an average logCPM below -1 were filtered out.
  2518. Each remaining gene was tested for differential abundance with respect
  2519. to globin blocking (GB) using edgeR’s quasi-likelihod F-test, fitting a
  2520. negative binomial generalized linear model to table of read counts in each
  2521. library.
  2522. For each gene, edgeR reported average abundance (logCPM),
  2523. \begin_inset Formula $\log_{2}$
  2524. \end_inset
  2525. fold change (logFC), p-value, and Benjamini-Hochberg adjusted false discovery
  2526. rate (FDR).
  2527. Each gene's logFC was plotted against its logCPM, colored by FDR.
  2528. Red points are significant at ≤10% FDR, and blue are not significant at
  2529. that threshold.
  2530. The alpha and beta globin genes targeted for blocking are marked with large
  2531. triangles, while all other genes are represented as small points.
  2532. \end_layout
  2533. \end_inset
  2534. \end_layout
  2535. \begin_layout Plain Layout
  2536. \end_layout
  2537. \end_inset
  2538. \end_layout
  2539. \begin_layout Standard
  2540. \begin_inset Flex TODO Note (inline)
  2541. status open
  2542. \begin_layout Plain Layout
  2543. Standardize on
  2544. \begin_inset Quotes eld
  2545. \end_inset
  2546. log2
  2547. \begin_inset Quotes erd
  2548. \end_inset
  2549. notation
  2550. \end_layout
  2551. \end_inset
  2552. \end_layout
  2553. \begin_layout Standard
  2554. The data do indeed show small systematic perturbations in gene levels (Figure
  2555. \begin_inset CommandInset ref
  2556. LatexCommand ref
  2557. reference "fig:MA-plot"
  2558. plural "false"
  2559. caps "false"
  2560. noprefix "false"
  2561. \end_inset
  2562. ).
  2563. Other than the 3 designated alpha and beta globin genes, two other genes
  2564. stand out as having especially large negative log fold changes: HBD and
  2565. LOC1021365.
  2566. HBD, delta globin, is most likely targeted by the blocking oligos due to
  2567. high sequence homology with the other globin genes.
  2568. LOC1021365 is the aforementioned ncRNA that is reverse-complementary to
  2569. one of the alpha-like genes and that would be expected to be removed during
  2570. the globin blocking step.
  2571. All other genes appear in a cluster centered vertically at 0, and the vast
  2572. majority of genes in this cluster show an absolute log2(FC) of 0.5 or less.
  2573. Nevertheless, many of these small perturbations are still statistically
  2574. significant, indicating that the globin blocking oligos likely cause very
  2575. small but non-zero systematic perturbations in measured gene expression
  2576. levels.
  2577. \end_layout
  2578. \begin_layout Standard
  2579. \begin_inset Float figure
  2580. wide false
  2581. sideways false
  2582. status open
  2583. \begin_layout Plain Layout
  2584. \align center
  2585. \begin_inset Graphics
  2586. filename graphics/Globin Paper/figure5 - corrplot.pdf
  2587. \end_inset
  2588. \end_layout
  2589. \begin_layout Plain Layout
  2590. \begin_inset Caption Standard
  2591. \begin_layout Plain Layout
  2592. \series bold
  2593. \begin_inset Argument 1
  2594. status collapsed
  2595. \begin_layout Plain Layout
  2596. Comparison of inter-sample gene abundance correlations with and without
  2597. globin blocking.
  2598. \end_layout
  2599. \end_inset
  2600. \begin_inset CommandInset label
  2601. LatexCommand label
  2602. name "fig:gene-abundance-correlations"
  2603. \end_inset
  2604. Comparison of inter-sample gene abundance correlations with and without
  2605. globin blocking (GB).
  2606. \series default
  2607. All libraries were normalized together as described in Figure 2, and genes
  2608. with an average abundance (logCPM, log2 counts per million reads counted)
  2609. less than -1 were filtered out.
  2610. Each gene’s logCPM was computed in each library using the edgeR cpm function.
  2611. For each pair of biological samples, the Pearson correlation between those
  2612. samples' GB libraries was plotted against the correlation between the same
  2613. samples’ non-GB libraries.
  2614. Each point represents an unique pair of samples.
  2615. The solid gray line shows a quantile-quantile plot of distribution of GB
  2616. correlations vs.
  2617. that of non-GB correlations.
  2618. The thin dashed line is the identity line, provided for reference.
  2619. \end_layout
  2620. \end_inset
  2621. \end_layout
  2622. \begin_layout Plain Layout
  2623. \end_layout
  2624. \end_inset
  2625. \end_layout
  2626. \begin_layout Standard
  2627. To evaluate the possibility of globin blocking causing random perturbations
  2628. and reducing sample quality, we computed the Pearson correlation between
  2629. logCPM values for every pair of samples with and without GB and plotted
  2630. them against each other (Figure
  2631. \begin_inset CommandInset ref
  2632. LatexCommand ref
  2633. reference "fig:gene-abundance-correlations"
  2634. plural "false"
  2635. caps "false"
  2636. noprefix "false"
  2637. \end_inset
  2638. ).
  2639. The plot indicated that the GB libraries have higher sample-to-sample correlati
  2640. ons than the non-GB libraries.
  2641. Parametric and nonparametric tests for differences between the correlations
  2642. with and without GB both confirmed that this difference was highly significant
  2643. (2-sided paired t-test: t = 37.2, df = 665, P ≪ 2.2e-16; 2-sided Wilcoxon
  2644. sign-rank test: V = 2195, P ≪ 2.2e-16).
  2645. Performing the same tests on the Spearman correlations gave the same conclusion
  2646. (t-test: t = 26.8, df = 665, P ≪ 2.2e-16; sign-rank test: V = 8781, P ≪ 2.2e-16).
  2647. The edgeR package was used to compute the overall biological coefficient
  2648. of variation (BCV) for GB and non-GB libraries, and found that globin blocking
  2649. resulted in a negligible increase in the BCV (0.417 with GB vs.
  2650. 0.400 without).
  2651. The near equality of the BCVs for both sets indicates that the higher correlati
  2652. ons in the GB libraries are most likely a result of the increased yield
  2653. of useful reads, which reduces the contribution of Poisson counting uncertainty
  2654. to the overall variance of the logCPM values
  2655. \begin_inset CommandInset citation
  2656. LatexCommand cite
  2657. key "McCarthy2012"
  2658. literal "false"
  2659. \end_inset
  2660. .
  2661. This improves the precision of expression measurements and more than offsets
  2662. the negligible increase in BCV.
  2663. \end_layout
  2664. \begin_layout Subsection*
  2665. More differentially expressed genes are detected with globin blocking
  2666. \end_layout
  2667. \begin_layout Standard
  2668. \begin_inset Float table
  2669. wide false
  2670. sideways false
  2671. status open
  2672. \begin_layout Plain Layout
  2673. \align center
  2674. \begin_inset Tabular
  2675. <lyxtabular version="3" rows="5" columns="5">
  2676. <features tabularvalignment="middle">
  2677. <column alignment="center" valignment="top">
  2678. <column alignment="center" valignment="top">
  2679. <column alignment="center" valignment="top">
  2680. <column alignment="center" valignment="top">
  2681. <column alignment="center" valignment="top">
  2682. <row>
  2683. <cell alignment="center" valignment="top" usebox="none">
  2684. \begin_inset Text
  2685. \begin_layout Plain Layout
  2686. \end_layout
  2687. \end_inset
  2688. </cell>
  2689. <cell alignment="center" valignment="top" usebox="none">
  2690. \begin_inset Text
  2691. \begin_layout Plain Layout
  2692. \end_layout
  2693. \end_inset
  2694. </cell>
  2695. <cell multicolumn="1" alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
  2696. \begin_inset Text
  2697. \begin_layout Plain Layout
  2698. \series bold
  2699. No Globin Blocking
  2700. \end_layout
  2701. \end_inset
  2702. </cell>
  2703. <cell multicolumn="2" alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
  2704. \begin_inset Text
  2705. \begin_layout Plain Layout
  2706. \end_layout
  2707. \end_inset
  2708. </cell>
  2709. <cell multicolumn="2" alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" rightline="true" usebox="none">
  2710. \begin_inset Text
  2711. \begin_layout Plain Layout
  2712. \end_layout
  2713. \end_inset
  2714. </cell>
  2715. </row>
  2716. <row>
  2717. <cell alignment="center" valignment="top" usebox="none">
  2718. \begin_inset Text
  2719. \begin_layout Plain Layout
  2720. \end_layout
  2721. \end_inset
  2722. </cell>
  2723. <cell alignment="center" valignment="top" usebox="none">
  2724. \begin_inset Text
  2725. \begin_layout Plain Layout
  2726. \end_layout
  2727. \end_inset
  2728. </cell>
  2729. <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  2730. \begin_inset Text
  2731. \begin_layout Plain Layout
  2732. \series bold
  2733. Up
  2734. \end_layout
  2735. \end_inset
  2736. </cell>
  2737. <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  2738. \begin_inset Text
  2739. \begin_layout Plain Layout
  2740. \series bold
  2741. NS
  2742. \end_layout
  2743. \end_inset
  2744. </cell>
  2745. <cell alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
  2746. \begin_inset Text
  2747. \begin_layout Plain Layout
  2748. \series bold
  2749. Down
  2750. \end_layout
  2751. \end_inset
  2752. </cell>
  2753. </row>
  2754. <row>
  2755. <cell multirow="3" alignment="center" valignment="middle" topline="true" bottomline="true" leftline="true" usebox="none">
  2756. \begin_inset Text
  2757. \begin_layout Plain Layout
  2758. \series bold
  2759. Globin-Blocking
  2760. \end_layout
  2761. \end_inset
  2762. </cell>
  2763. <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  2764. \begin_inset Text
  2765. \begin_layout Plain Layout
  2766. \series bold
  2767. Up
  2768. \end_layout
  2769. \end_inset
  2770. </cell>
  2771. <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  2772. \begin_inset Text
  2773. \begin_layout Plain Layout
  2774. \family roman
  2775. \series medium
  2776. \shape up
  2777. \size normal
  2778. \emph off
  2779. \bar no
  2780. \strikeout off
  2781. \xout off
  2782. \uuline off
  2783. \uwave off
  2784. \noun off
  2785. \color none
  2786. 231
  2787. \end_layout
  2788. \end_inset
  2789. </cell>
  2790. <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  2791. \begin_inset Text
  2792. \begin_layout Plain Layout
  2793. \family roman
  2794. \series medium
  2795. \shape up
  2796. \size normal
  2797. \emph off
  2798. \bar no
  2799. \strikeout off
  2800. \xout off
  2801. \uuline off
  2802. \uwave off
  2803. \noun off
  2804. \color none
  2805. 515
  2806. \end_layout
  2807. \end_inset
  2808. </cell>
  2809. <cell alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
  2810. \begin_inset Text
  2811. \begin_layout Plain Layout
  2812. \family roman
  2813. \series medium
  2814. \shape up
  2815. \size normal
  2816. \emph off
  2817. \bar no
  2818. \strikeout off
  2819. \xout off
  2820. \uuline off
  2821. \uwave off
  2822. \noun off
  2823. \color none
  2824. 2
  2825. \end_layout
  2826. \end_inset
  2827. </cell>
  2828. </row>
  2829. <row>
  2830. <cell multirow="4" alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  2831. \begin_inset Text
  2832. \begin_layout Plain Layout
  2833. \end_layout
  2834. \end_inset
  2835. </cell>
  2836. <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  2837. \begin_inset Text
  2838. \begin_layout Plain Layout
  2839. \series bold
  2840. NS
  2841. \end_layout
  2842. \end_inset
  2843. </cell>
  2844. <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  2845. \begin_inset Text
  2846. \begin_layout Plain Layout
  2847. \family roman
  2848. \series medium
  2849. \shape up
  2850. \size normal
  2851. \emph off
  2852. \bar no
  2853. \strikeout off
  2854. \xout off
  2855. \uuline off
  2856. \uwave off
  2857. \noun off
  2858. \color none
  2859. 160
  2860. \end_layout
  2861. \end_inset
  2862. </cell>
  2863. <cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
  2864. \begin_inset Text
  2865. \begin_layout Plain Layout
  2866. \family roman
  2867. \series medium
  2868. \shape up
  2869. \size normal
  2870. \emph off
  2871. \bar no
  2872. \strikeout off
  2873. \xout off
  2874. \uuline off
  2875. \uwave off
  2876. \noun off
  2877. \color none
  2878. 11235
  2879. \end_layout
  2880. \end_inset
  2881. </cell>
  2882. <cell alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
  2883. \begin_inset Text
  2884. \begin_layout Plain Layout
  2885. \family roman
  2886. \series medium
  2887. \shape up
  2888. \size normal
  2889. \emph off
  2890. \bar no
  2891. \strikeout off
  2892. \xout off
  2893. \uuline off
  2894. \uwave off
  2895. \noun off
  2896. \color none
  2897. 136
  2898. \end_layout
  2899. \end_inset
  2900. </cell>
  2901. </row>
  2902. <row>
  2903. <cell multirow="4" alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
  2904. \begin_inset Text
  2905. \begin_layout Plain Layout
  2906. \end_layout
  2907. \end_inset
  2908. </cell>
  2909. <cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
  2910. \begin_inset Text
  2911. \begin_layout Plain Layout
  2912. \series bold
  2913. Down
  2914. \end_layout
  2915. \end_inset
  2916. </cell>
  2917. <cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
  2918. \begin_inset Text
  2919. \begin_layout Plain Layout
  2920. \family roman
  2921. \series medium
  2922. \shape up
  2923. \size normal
  2924. \emph off
  2925. \bar no
  2926. \strikeout off
  2927. \xout off
  2928. \uuline off
  2929. \uwave off
  2930. \noun off
  2931. \color none
  2932. 0
  2933. \end_layout
  2934. \end_inset
  2935. </cell>
  2936. <cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
  2937. \begin_inset Text
  2938. \begin_layout Plain Layout
  2939. \family roman
  2940. \series medium
  2941. \shape up
  2942. \size normal
  2943. \emph off
  2944. \bar no
  2945. \strikeout off
  2946. \xout off
  2947. \uuline off
  2948. \uwave off
  2949. \noun off
  2950. \color none
  2951. 548
  2952. \end_layout
  2953. \end_inset
  2954. </cell>
  2955. <cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" rightline="true" usebox="none">
  2956. \begin_inset Text
  2957. \begin_layout Plain Layout
  2958. \family roman
  2959. \series medium
  2960. \shape up
  2961. \size normal
  2962. \emph off
  2963. \bar no
  2964. \strikeout off
  2965. \xout off
  2966. \uuline off
  2967. \uwave off
  2968. \noun off
  2969. \color none
  2970. 127
  2971. \end_layout
  2972. \end_inset
  2973. </cell>
  2974. </row>
  2975. </lyxtabular>
  2976. \end_inset
  2977. \end_layout
  2978. \begin_layout Plain Layout
  2979. \begin_inset Caption Standard
  2980. \begin_layout Plain Layout
  2981. \series bold
  2982. \begin_inset Argument 1
  2983. status open
  2984. \begin_layout Plain Layout
  2985. Comparison of significantly differentially expressed genes with and without
  2986. globin blocking.
  2987. \end_layout
  2988. \end_inset
  2989. \begin_inset CommandInset label
  2990. LatexCommand label
  2991. name "tab:Comparison-of-significant"
  2992. \end_inset
  2993. Comparison of significantly differentially expressed genes with and without
  2994. globin blocking.
  2995. \series default
  2996. Up, Down: Genes significantly up/down-regulated in post-transplant samples
  2997. relative to pre-transplant samples, with a false discovery rate of 10%
  2998. or less.
  2999. NS: Non-significant genes (false discovery rate greater than 10%).
  3000. \end_layout
  3001. \end_inset
  3002. \end_layout
  3003. \begin_layout Plain Layout
  3004. \end_layout
  3005. \end_inset
  3006. \end_layout
  3007. \begin_layout Standard
  3008. To compare performance on differential gene expression tests, we took subsets
  3009. of both the GB and non-GB libraries with exactly one pre-transplant and
  3010. one post-transplant sample for each animal that had paired samples available
  3011. for analysis (N=7 animals, N=14 samples in each subset).
  3012. The same test for pre- vs.
  3013. post-transplant differential gene expression was performed on the same
  3014. 7 pairs of samples from GB libraries and non-GB libraries, in each case
  3015. using an FDR of 10% as the threshold of significance.
  3016. Out of 12954 genes that passed the detection threshold in both subsets,
  3017. 358 were called significantly differentially expressed in the same direction
  3018. in both sets; 1063 were differentially expressed in the GB set only; 296
  3019. were differentially expressed in the non-GB set only; 2 genes were called
  3020. significantly up in the GB set but significantly down in the non-GB set;
  3021. and the remaining 11235 were not called differentially expressed in either
  3022. set.
  3023. These data are summarized in Table
  3024. \begin_inset CommandInset ref
  3025. LatexCommand ref
  3026. reference "tab:Comparison-of-significant"
  3027. plural "false"
  3028. caps "false"
  3029. noprefix "false"
  3030. \end_inset
  3031. .
  3032. The differences in BCV calculated by EdgeR for these subsets of samples
  3033. were negligible (BCV = 0.302 for GB and 0.297 for non-GB).
  3034. \end_layout
  3035. \begin_layout Standard
  3036. The key point is that the GB data results in substantially more differentially
  3037. expressed calls than the non-GB data.
  3038. Since there is no gold standard for this dataset, it is impossible to be
  3039. certain whether this is due to under-calling of differential expression
  3040. in the non-GB samples or over-calling in the GB samples.
  3041. However, given that both datasets are derived from the same biological
  3042. samples and have nearly equal BCVs, it is more likely that the larger number
  3043. of DE calls in the GB samples are genuine detections that were enabled
  3044. by the higher sequencing depth and measurement precision of the GB samples.
  3045. Note that the same set of genes was considered in both subsets, so the
  3046. larger number of differentially expressed gene calls in the GB data set
  3047. reflects a greater sensitivity to detect significant differential gene
  3048. expression and not simply the larger total number of detected genes in
  3049. GB samples described earlier.
  3050. \end_layout
  3051. \begin_layout Section
  3052. Discussion
  3053. \end_layout
  3054. \begin_layout Standard
  3055. The original experience with whole blood gene expression profiling on DNA
  3056. microarrays demonstrated that the high concentration of globin transcripts
  3057. reduced the sensitivity to detect genes with relatively low expression
  3058. levels, in effect, significantly reducing the sensitivity.
  3059. To address this limitation, commercial protocols for globin reduction were
  3060. developed based on strategies to block globin transcript amplification
  3061. during labeling or physically removing globin transcripts by affinity bead
  3062. methods
  3063. \begin_inset CommandInset citation
  3064. LatexCommand cite
  3065. key "Winn2010"
  3066. literal "false"
  3067. \end_inset
  3068. .
  3069. More recently, using the latest generation of labeling protocols and arrays,
  3070. it was determined that globin reduction was no longer necessary to obtain
  3071. sufficient sensitivity to detect differential transcript expression
  3072. \begin_inset CommandInset citation
  3073. LatexCommand cite
  3074. key "NuGEN2010"
  3075. literal "false"
  3076. \end_inset
  3077. .
  3078. However, we are not aware of any publications using these currently available
  3079. protocols the with latest generation of microarrays that actually compare
  3080. the detection sensitivity with and without globin reduction.
  3081. However, in practice this has now been adopted generally primarily driven
  3082. by concerns for cost control.
  3083. The main objective of our work was to directly test the impact of globin
  3084. gene transcripts and a new globin blocking protocol for application to
  3085. the newest generation of differential gene expression profiling determined
  3086. using next generation sequencing.
  3087. \end_layout
  3088. \begin_layout Standard
  3089. The challenge of doing global gene expression profiling in cynomolgus monkeys
  3090. is that the current available arrays were never designed to comprehensively
  3091. cover this genome and have not been updated since the first assemblies
  3092. of the cynomolgus genome were published.
  3093. Therefore, we determined that the best strategy for peripheral blood profiling
  3094. was to do deep RNA-seq and inform the workflow using the latest available
  3095. genome assembly and annotation
  3096. \begin_inset CommandInset citation
  3097. LatexCommand cite
  3098. key "Wilson2013"
  3099. literal "false"
  3100. \end_inset
  3101. .
  3102. However, it was not immediately clear whether globin reduction was necessary
  3103. for RNA-seq or how much improvement in efficiency or sensitivity to detect
  3104. differential gene expression would be achieved for the added cost and work.
  3105. \end_layout
  3106. \begin_layout Standard
  3107. We only found one report that demonstrated that globin reduction significantly
  3108. improved the effective read yields for sequencing of human peripheral blood
  3109. cell RNA using a DeepSAGE protocol
  3110. \begin_inset CommandInset citation
  3111. LatexCommand cite
  3112. key "Mastrokolias2012"
  3113. literal "false"
  3114. \end_inset
  3115. .
  3116. The approach to DeepSAGE involves two different restriction enzymes that
  3117. purify and then tag small fragments of transcripts at specific locations
  3118. and thus, significantly reduces the complexity of the transcriptome.
  3119. Therefore, we could not determine how DeepSAGE results would translate
  3120. to the common strategy in the field for assaying the entire transcript
  3121. population by whole-transcriptome 3’-end RNA-seq.
  3122. Furthermore, if globin reduction is necessary, we also needed a globin
  3123. reduction method specific to cynomolgus globin sequences that would work
  3124. an organism for which no kit is available off the shelf.
  3125. \end_layout
  3126. \begin_layout Standard
  3127. As mentioned above, the addition of globin blocking oligos has a very small
  3128. impact on measured expression levels of gene expression.
  3129. However, this is a non-issue for the purposes of differential expression
  3130. testing, since a systematic change in a gene in all samples does not affect
  3131. relative expression levels between samples.
  3132. However, we must acknowledge that simple comparisons of gene expression
  3133. data obtained by GB and non-GB protocols are not possible without additional
  3134. normalization.
  3135. \end_layout
  3136. \begin_layout Standard
  3137. More importantly, globin blocking not only nearly doubles the yield of usable
  3138. reads, it also increases inter-sample correlation and sensitivity to detect
  3139. differential gene expression relative to the same set of samples profiled
  3140. without blocking.
  3141. In addition, globin blocking does not add a significant amount of random
  3142. noise to the data.
  3143. Globin blocking thus represents a cost-effective way to squeeze more data
  3144. and statistical power out of the same blood samples and the same amount
  3145. of sequencing.
  3146. In conclusion, globin reduction greatly increases the yield of useful RNA-seq
  3147. reads mapping to the rest of the genome, with minimal perturbations in
  3148. the relative levels of non-globin genes.
  3149. Based on these results, globin transcript reduction using sequence-specific,
  3150. complementary blocking oligonucleotides is recommended for all deep RNA-seq
  3151. of cynomolgus and other nonhuman primate blood samples.
  3152. \end_layout
  3153. \begin_layout Chapter
  3154. Future Directions
  3155. \end_layout
  3156. \begin_layout Itemize
  3157. Study other epigenetic marks in more contexts
  3158. \end_layout
  3159. \begin_deeper
  3160. \begin_layout Itemize
  3161. DNA methylation, histone marks, chromatin accessibility & conformation in
  3162. CD4 T-cells
  3163. \end_layout
  3164. \begin_layout Itemize
  3165. Also look at other types lymphocytes: CD8 T-cells, B-cells, NK cells
  3166. \end_layout
  3167. \end_deeper
  3168. \begin_layout Itemize
  3169. Investigate epigenetic regulation of lifespan extension in
  3170. \emph on
  3171. C.
  3172. elegans
  3173. \end_layout
  3174. \begin_deeper
  3175. \begin_layout Itemize
  3176. ChIP-seq of important transcriptional regulators to see how transcriptional
  3177. drift is prevented
  3178. \end_layout
  3179. \end_deeper
  3180. \begin_layout Standard
  3181. \begin_inset ERT
  3182. status open
  3183. \begin_layout Plain Layout
  3184. % Use "References" instead of "Bibliography"
  3185. \end_layout
  3186. \begin_layout Plain Layout
  3187. \backslash
  3188. renewcommand{
  3189. \backslash
  3190. bibname}{References}
  3191. \end_layout
  3192. \end_inset
  3193. \end_layout
  3194. \begin_layout Standard
  3195. \begin_inset Flex TODO Note (inline)
  3196. status open
  3197. \begin_layout Plain Layout
  3198. Check bib entry formatting & sort order
  3199. \end_layout
  3200. \end_inset
  3201. \end_layout
  3202. \begin_layout Standard
  3203. \begin_inset CommandInset bibtex
  3204. LatexCommand bibtex
  3205. btprint "btPrintCited"
  3206. bibfiles "refs"
  3207. options "bibtotoc,unsrt"
  3208. \end_inset
  3209. \end_layout
  3210. \end_body
  3211. \end_document