"normalize.quantiles.robust" produces the SAME intensities from different weights?!
1
0
Entering edit mode
k. brand ▴ 420
@k-brand-1874
Last seen 9.6 years ago
Dear BioCers, My interpretation of the function "normalize.quantiles.robust" is the ability to selectively increase or reduce the contribution of specific chips during normalization. Although i get different results from equally weighted chips (which is the same as "justRMA" as a verification), all variations of unequal weightings produce do the SAME intensities?! See 2 examples below. Can some one find the error in my script, or better yet, give an example of a better way to achieve my objective- to selectively 'pull' low intensities of duplicate chips form a poor hyb (on expired arrays), up to the levels of duplicate chips of a good hyb(on unexpired arrays). TIA, Karl >### "weight_is_1&10.R" > > setwd("D:/brand 8/CORE-SHELL/Diff methods/ID T0 all/normalisations/Quantiles-robust") > library(affy) > dat <- ReadAffy() > # list with weights for robust normalization > param <- list(weights=c(1,1,10,10,1,1,10,10,1,1,10,10)) > # performs robust quantile normalization > # summarization is performed using medianpolish > eset <- expresso(dat, bgcorrect.method="rma", + normalize.method="quantiles.robust", normalize.param=param, + pmcorrect.method="pmonly", summary.method = "medianpolish") background correction: rma normalization: quantiles.robust PM/MM correction : pmonly expression values: medianpolish background correcting...done. normalizing...Chip weights are 1 1 10 10 1 1 10 10 1 1 10 10 Chip weights are 1 1 10 10 1 1 10 10 1 1 10 10 done. 45101 ids to be processed | | |####################| > exprs2excel(eset, "weight_is_1&10.csv") > for probset 1415670_at 6.268887235 5.946937963 7.191120262 7.087367938 6.145685893 6.095744472 7.473473796 7.21058045 6.105367399 6.130141537 7.324310607 7.059578097 > ### "weight_is_10&1.R" > dat <- ReadAffy() > # list with weights for robust normalization > param <- list(weights=c(10,10,1,1,10,10,1,1,10,10,1,1)) > # performs robust quantile normalization > # summarization is performed using medianpolish > eset <- expresso(dat, bgcorrect.method="rma", + normalize.method="quantiles.robust", normalize.param=param, + pmcorrect.method="pmonly", summary.method = "medianpolish") background correction: rma normalization: quantiles.robust PM/MM correction : pmonly expression values: medianpolish background correcting...done. normalizing...Chip weights are 10 10 1 1 10 10 1 1 10 10 1 1 Chip weights are 10 10 1 1 10 10 1 1 10 10 1 1 done. 45101 ids to be processed | | |####################| > exprs2excel(eset, "weight_is_10&1.csv") for probset 1415670_at 6.268887235 5.946937963 7.191120262 7.087367938 6.145685893 6.095744472 7.473473796 7.21058045 6.105367399 6.130141537 7.324310607 7.059578097 ============================================================== > sessionInfo() Version 2.3.0 (2006-04-24) i386-pc-mingw32 attached base packages: [1] "tools" "methods" "stats" "graphics" "grDevices" "utils" [7] "datasets" "base" other attached packages: affyPLM gcrma matchprobes affydata mouse4302cdf affy "1.8.0" "2.4.1" "1.4.0" "1.8.0" "1.12.0" "1.10.0" affyio Biobase "1.0.0" "1.10.0" -- Karl Brand <k.brand at="" erasmusmc.nl=""> Department of Cell Biology and Genetics Erasmus MC Dr Molewaterplein 50 3015 GE Rotterdam lab +31 (0)10 408 7409 fax +31 (0)10 408 9468
Normalization affydata gcrma matchprobes affyPLM affyio Normalization affydata gcrma • 1.1k views
ADD COMMENT
0
Entering edit mode
Ben Bolstad ★ 1.2k
@ben-bolstad-1494
Last seen 6.6 years ago
Thanks for this report. It turned out that there was a minor bug in normalize.quantiles.robust() and incorrect parameters were being passed down to the C code. This is due to some changes I made many months ago now. A corrected version of the function will appear in the next version of affy. But briefly the final line in that function should have been: .Call("R_qnorm_robust_c",x,copy,weights,as.integer(use.median),as.int eger(use.log2),as.integer(use.huber),PACKAGE="affy") Best, Ben On Wed, 2006-09-20 at 17:29 +0200, k. brand wrote: > Dear BioCers, > > My interpretation of the function "normalize.quantiles.robust" is the > ability to selectively increase or reduce the contribution of specific > chips during normalization. > > Although i get different results from equally weighted chips (which is > the same as "justRMA" as a verification), all variations of unequal > weightings produce do the SAME intensities?! See 2 examples below. > > Can some one find the error in my script, or better yet, give an example > of a better way to achieve my objective- > > to selectively 'pull' low intensities of duplicate chips form a poor hyb > (on expired arrays), up to the levels of duplicate chips of a good > hyb(on unexpired arrays). > > TIA, > > Karl > >
ADD COMMENT
0
Entering edit mode
Ben, thank you for the fast help- it worked! Karl on 9/21/2006 5:16 AM Ben Bolstad said the following: > Thanks for this report. It turned out that there was a minor bug in > normalize.quantiles.robust() and incorrect parameters were being passed > down to the C code. This is due to some changes I made many months ago > now. > > A corrected version of the function will appear in the next version of > affy. But briefly the final line in that function should have been: > > > .Call("R_qnorm_robust_c",x,copy,weights,as.integer(use.median),as.i nteger(use.log2),as.integer(use.huber),PACKAGE="affy") > > Best, > > Ben > > On Wed, 2006-09-20 at 17:29 +0200, k. brand wrote: >> Dear BioCers, >> >> My interpretation of the function "normalize.quantiles.robust" is the >> ability to selectively increase or reduce the contribution of specific >> chips during normalization. >> >> Although i get different results from equally weighted chips (which is >> the same as "justRMA" as a verification), all variations of unequal >> weightings produce do the SAME intensities?! See 2 examples below. >> >> Can some one find the error in my script, or better yet, give an example >> of a better way to achieve my objective- >> >> to selectively 'pull' low intensities of duplicate chips form a poor hyb >> (on expired arrays), up to the levels of duplicate chips of a good >> hyb(on unexpired arrays). >> >> TIA, >> >> Karl >> >> > > -- Karl Brand <k.brand at="" erasmusmc.nl=""> Department of Genetics Erasmus MC Dr Molewaterplein 50 3015 GE Rotterdam lab +31 (0)10 408 7409 fax +31 (0)10 408 9468
ADD REPLY

Login before adding your answer.

Traffic: 595 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6