coefficient not estimable (limma warning)
2
0
Entering edit mode
Natasha ▴ 440
@natasha-4640
Last seen 7.8 years ago
Dear List, I do know that this question has been approached before, but am not sure about the reliability of the list of genes. I have data from an experiment that was run in two batches (3 groups per batch, 4 samples per group, 2 array-chips per bacth), these need to be combined and analysed. Thus there are 6 groups, and I know from the QC that there is a strong batch effect, which I need to add to the model. However when I run the analysis it says batch is not estimable. I am unsure now whether the limma list of differentially expressed genes is valid or not, as it might have not accounted for batch variation. (I do appreciate that it would also be a confounding factor, so one would not be certain of true differential effects without further experimental verification, but I am still uncertain of my list of genes). Code below: design2 = model.matrix(~0+group+batch) design2 WT KO H1KO H2KO TripleKO P3 batch2 1 0 1 0 0 0 0 0 2 0 1 0 0 0 0 0 3 0 1 0 0 0 0 0 4 0 1 0 0 0 0 0 5 0 0 1 0 0 0 0 6 0 0 1 0 0 0 0 7 0 0 1 0 0 0 0 8 0 0 1 0 0 0 0 9 1 0 0 0 0 0 0 10 1 0 0 0 0 0 0 11 1 0 0 0 0 0 0 12 1 0 0 0 0 0 0 13 0 0 0 0 1 0 1 14 0 0 0 0 1 0 1 15 0 0 0 0 1 0 1 16 0 0 0 0 1 0 1 17 0 0 0 1 0 0 1 18 0 0 0 1 0 0 1 19 0 0 0 1 0 0 1 20 0 0 0 1 0 0 1 21 0 0 0 0 0 1 1 22 0 0 0 0 0 1 1 23 0 0 0 0 0 1 1 24 0 0 0 0 0 1 1 nonEstimable(design2) [1] "batch2" fit2 <- lmFit(sig.norm, design2) Coefficients not estimable: batch2 Warning message: Partial NA coefficients for 26084 probe(s) con.mat2 = makeContrasts(KO-WT, H1KO-WT, H2KO-WT, TripleKO-WT, P3-WT, P3-KO, TripleKO-KO, levels=design2) fit2b <- contrasts.fit(fit2, con.mat2) fit2b <- eBayes(fit2b) Any help and suggestions would be much appreciated. Many Thanks, Natasha
limma limma • 7.7k views
ADD COMMENT
1
Entering edit mode
@moshe-olshansky-4491
Last seen 7.8 years ago
Hi Natasha, In your case, H2KO + TripleKO + P3 = batch2 (column-wise), so your matrix's columns are dependent and this causes the technical problem. Intuitively, since you have 3 samples in batch 1 and 3 other samples in batch 2, you can never know whether you have a batch effect or your one set of samples is very different from the other one. Best regards, Moshe. > Dear List, > > I do know that this question has been approached before, but am not sure > about > the reliability of the list of genes. > > I have data from an experiment that was run in two batches (3 groups per > batch, > 4 samples per group, 2 array-chips per bacth), these need to be combined > and > analysed. > > Thus there are 6 groups, and I know from the QC that there is a strong > batch > effect, which I need to add to the model. However when I run the analysis > it > says batch is not estimable. > > I am unsure now whether the limma list of differentially expressed genes > is > valid or not, as it might have not accounted for batch variation. (I do > appreciate that it would also be a confounding factor, so one would not be > certain of true differential effects without further experimental > verification, > but I am still uncertain of my list of genes). > > Code below: > > design2 = model.matrix(~0+group+batch) > design2 > WT KO H1KO H2KO TripleKO P3 batch2 > 1 0 1 0 0 0 0 0 > 2 0 1 0 0 0 0 0 > 3 0 1 0 0 0 0 0 > 4 0 1 0 0 0 0 0 > 5 0 0 1 0 0 0 0 > 6 0 0 1 0 0 0 0 > 7 0 0 1 0 0 0 0 > 8 0 0 1 0 0 0 0 > 9 1 0 0 0 0 0 0 > 10 1 0 0 0 0 0 0 > 11 1 0 0 0 0 0 0 > 12 1 0 0 0 0 0 0 > 13 0 0 0 0 1 0 1 > 14 0 0 0 0 1 0 1 > 15 0 0 0 0 1 0 1 > 16 0 0 0 0 1 0 1 > 17 0 0 0 1 0 0 1 > 18 0 0 0 1 0 0 1 > 19 0 0 0 1 0 0 1 > 20 0 0 0 1 0 0 1 > 21 0 0 0 0 0 1 1 > 22 0 0 0 0 0 1 1 > 23 0 0 0 0 0 1 1 > 24 0 0 0 0 0 1 1 > > nonEstimable(design2) > [1] "batch2" > > fit2 <- lmFit(sig.norm, design2) > Coefficients not estimable: batch2 > Warning message: > Partial NA coefficients for 26084 probe(s) > > con.mat2 = makeContrasts(KO-WT, H1KO-WT, H2KO-WT, TripleKO-WT, P3-WT, > P3-KO, > TripleKO-KO, levels=design2) > > fit2b <- contrasts.fit(fit2, con.mat2) > fit2b <- eBayes(fit2b) > > > Any help and suggestions would be much appreciated. > > Many Thanks, > Natasha > > _______________________________________________ > Bioconductor mailing list > Bioconductor at r-project.org > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: > http://news.gmane.org/gmane.science.biology.informatics.conductor > -- Moshe Olshansky Division of Bioinformatics The Walter & Eliza Hall Institute of Medical Research 1G Royal Parade, Parkville, Vic 3052 e-mail: olshansky at wehi.edu.au tel: (03) 9345 2697 ______________________________________________________________________ The information in this email is confidential and intend...{{dropped:4}}
ADD COMMENT
1
Entering edit mode
Naomi Altman ★ 6.0k
@naomi-altman-380
Last seen 14 months ago
United States
According to your design matrix, the first 3 treatments were run in batch 0 and the last 3 in batch 1. So batch is confounded with treatment. If so, there is no way to analyze the experiment that produces a valid comparison between treatments run in the different batches. Naomi At 10:17 AM 5/11/2011, Natasha wrote: >Dear List, > >I do know that this question has been approached before, but am not sure about >the reliability of the list of genes. > >I have data from an experiment that was run in two batches (3 groups >per batch, >4 samples per group, 2 array-chips per bacth), these need to be combined and >analysed. > >Thus there are 6 groups, and I know from the QC that there is a strong batch >effect, which I need to add to the model. However when I run the analysis it >says batch is not estimable. > >I am unsure now whether the limma list of differentially expressed genes is >valid or not, as it might have not accounted for batch variation. (I do >appreciate that it would also be a confounding factor, so one would not be >certain of true differential effects without further experimental >verification, >but I am still uncertain of my list of genes). > >Code below: > >design2 = model.matrix(~0+group+batch) >design2 > WT KO H1KO H2KO TripleKO P3 batch2 >1 0 1 0 0 0 0 0 >2 0 1 0 0 0 0 0 >3 0 1 0 0 0 0 0 >4 0 1 0 0 0 0 0 >5 0 0 1 0 0 0 0 >6 0 0 1 0 0 0 0 >7 0 0 1 0 0 0 0 >8 0 0 1 0 0 0 0 >9 1 0 0 0 0 0 0 >10 1 0 0 0 0 0 0 >11 1 0 0 0 0 0 0 >12 1 0 0 0 0 0 0 >13 0 0 0 0 1 0 1 >14 0 0 0 0 1 0 1 >15 0 0 0 0 1 0 1 >16 0 0 0 0 1 0 1 >17 0 0 0 1 0 0 1 >18 0 0 0 1 0 0 1 >19 0 0 0 1 0 0 1 >20 0 0 0 1 0 0 1 >21 0 0 0 0 0 1 1 >22 0 0 0 0 0 1 1 >23 0 0 0 0 0 1 1 >24 0 0 0 0 0 1 1 > >nonEstimable(design2) > [1] "batch2" > >fit2 <- lmFit(sig.norm, design2) > Coefficients not estimable: batch2 > Warning message: > Partial NA coefficients for 26084 probe(s) > >con.mat2 = makeContrasts(KO-WT, H1KO-WT, H2KO-WT, TripleKO-WT, P3-WT, P3-KO, >TripleKO-KO, levels=design2) > >fit2b <- contrasts.fit(fit2, con.mat2) >fit2b <- eBayes(fit2b) > > >Any help and suggestions would be much appreciated. > >Many Thanks, >Natasha > >_______________________________________________ >Bioconductor mailing list >Bioconductor at r-project.org >https://stat.ethz.ch/mailman/listinfo/bioconductor >Search the archives: >http://news.gmane.org/gmane.science.biology.informatics.conductor
ADD COMMENT

Login before adding your answer.

Traffic: 253 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