[BioC] variation in output -using gcrma: bg correction

[James W. MacDonald]

Sucheta Tripathy wrote:
> Dear Group
> 
> when I run background(using GCRMA) correction for  chips:
> 
> "10_1331.1I.R208L.CEL"  "9_C101.1F.R9F2L.CEL" and for chips
> "10_1195.1E.R119U.CEL"  "9_C101.1F.R9F2L.CEL" keeping
> 9_C101.1F.R9F2L.CEL as constant, For both the runs I get slightly
> different outputs for chip 9_C101.1F.R9F2L.CEL (Data below). This does not
> happen when GSB.adjust is set to FALSE. Also we have noticed the variation
> seem to get lesser when we try more chips together.Any reason why it
> should be happening this way?

I would recommend reading (or re-reading) the GCRMA papers by Wu and 
Irizarry that can be found on Rafael Irizarry's website. The question 
you pose indicates a fundamental lack of knowledge about the procedure 
you are using, which can be dangerous, especially since you appear to 
want to do some non-standard things.

Anyway, the short answer is that the background is based on a model fit 
using the MM probes (from all chips in the data set). If you vary what 
chips you are using, by definition the background estimates will change. 
In addition, as with all models, the more data you have, the better and 
more stable the estimates, so more chips == better and more stable 
background estimates.

Best,

Jim

> 
> Many thanks
> 
> Sucheta
> 
> My code:
> -------
>   obj<-ReadAffy(filenames=file)
>   tmp1<-bg.adjust.gcrma(obj,optical.correct=T,affinity.info=affinfo)
>   outfile<-paste(outdir, celfile[i,1], "_bg.txt", sep="")
>   write.table(pm(tmp1), file=outfile, row.names=FALSE, col.names=TRUE,
> sep="\t")
> ----------------------------------------------------------------------
> 
> 
> "10_1331.1I.R208L.CEL"  "9_C101.1F.R9F2L.CEL"
> ---------------------------------------------
> 12.5854077145271    4.27009199115734
> 135.577788495271    115.177321956904
> 113.171763014501    124.781182045557
> 2.69508615100678    13.3527712251500
> 14.5462397159524    36.7662738574386
> 94.6480456343775    48.972931999473
> 230.750229599019    212.534925776574
> 242.058765502072    337.251490520153
> 233.871977591185    339.413283090951
> 134.355125363486    108.047618775561
> 136.685235380017    126.239703801863
> 76.2625488684543    45.3623158856175
> 148.251139760089    120.315265011794
> 328.121529159736    319.366509787047
> 291.865380794849    361.992073209429
> 3.97218796139000    4.20902848830535
> 14.6419514021341    31.6450267847893
> 28.8699339528603    67.3237575282198
> 110.496037483661    111.806067290232
> 22.2876523362755    41.5481196151883
> 205.194483921088    276.995029856902
> 84.8559677803056    96.0077121491932
> 778.222646971625    559.863168099135
> 4.83202857118797    11.3149613102092
> 3.66166154528273    39.6717762417948
> 67.7761672656053    94.6394189279102
> 112.500673213605    190.985739398352
> 171.007645212454    200.035151512867
> 492.236603432973    352.584270678829
> 130.800353580651    155.324195356517
> 146.303381286542    160.442554666605
> 4.34069757710996    61.4777011118527
> 88.9169964732945    84.2320613854983
> 69.1622655457769    60.1255829409642
> 382.758342400691    429.238548268896
> 3.90087434250685    4.02923006445781
> 245.047319536046    186.199916603738
> 4.60503013970163    16.7179032017834
> 45.8974167676779    58.6091870023334
> 152.737507677336    87.4646339988696
> ===================================================
> 
> "10_1195.1E.R119U.CEL"  "9_C101.1F.R9F2L.CEL"
> ---------------------------------------------
> 34.9453663026134    4.35753552341506
> 120.903734762231    116.962904914229
> 111.293345773408    133.421207928372
> 2.96910233809292    14.4050686356057
> 34.7269027797637    35.8835240852081
> 85.3109066158579    45.2284068764267
> 222.081465596656    201.346914814921
> 378.991271676870    336.282180468649
> 324.355847430672    337.586327159885
> 148.252162519851    107.763579952789
> 146.760939437243    127.027138121574
> 68.9112298899552    45.5584187250637
> 178.300028533109    125.831740523771
> 457.092497175424    331.253094567131
> 349.374485133996    357.721065463285
> 8.39630366380937    4.30526177658911
> 50.2501294702745    31.1539961383794
> 51.2772385586404    66.8937884059372
> 169.473350033569    117.993461506227
> 47.5601861370703    41.0301951390256
> 238.568680512591    271.103448196271
> 122.668905036686    92.7991203760259
> 710.506330695352    568.473304793075
> 14.1193442910318    11.2664752571536
> 29.6858453280301    41.0915584209232
> 78.9811024872745    98.557906977124
> 143.137824860583    200.312672914717
> 228.952312559228    208.952150636992
> 414.93711559597 364.157958616779
> 193.573470074735    156.743939369768
> 151.710594739511    165.182915986475
> 16.3790389361525    62.1220611327726
> 63.6344501007354    81.4281142367824
> 52.6876908746722    58.4074520261634
> 356.890313613082    426.263965724419
> 4.0847500724139 4.13161532597561
> 233.32353554667 185.840713238473
> 4.71348339142599    16.7565881676473
> 89.4105877815801    55.7427731671095
> 162.930238504639    91.1574660327713
> 27.6720869896949    30.8456197258235
> 107.041950691333    17.5365838616139
> 132.491703246086    94.8472761626705
> 165.348305450273    118.806052283819
> 428.453296492502    438.259964898252
> 38.5832188166168    20.0643402347992
> 


-- 
James W. MacDonald, M.S.
Biostatistician
Affymetrix and cDNA Microarray Core
University of Michigan Cancer Center
1500 E. Medical Center Drive
7410 CCGC
Ann Arbor MI 48109
734-647-5623


**********************************************************
Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues.