Question: DESeq2 more differentially expressed genes
0
gravatar for lmf.bill@gmail.com
4.9 years ago by
United States
lmf.bill@gmail.com10 wrote:

I used DESeq2 and found it report more deferentially expressed genes than before. Actually, it is not reasonable.

I noticed the padj column  is different from the values I am calculating using the p.adjust (pvalue column). I would like you to tell me the incorrect of my calculation, or some bugs in DESeq2. Thanks a lot. In the following, I pasted my sessionInfo()

 

> sessionInfo()
R version 3.1.0 (2014-04-10)
Platform: x86_64-unknown-linux-gnu (64-bit)

locale:
[1] C

attached base packages:
[1] splines   parallel  stats     graphics  grDevices utils     datasets
[8] methods   base     

other attached packages:
 [1] cqn_1.10.0              quantreg_5.05           SparseM_1.05           
 [4] preprocessCore_1.26.1   nor1mix_1.2-0           mclust_4.3             
 [7] DESeq2_1.4.5            RcppArmadillo_0.4.400.0 Rcpp_0.11.2            
[10] GenomicRanges_1.16.4    GenomeInfoDb_1.0.2      IRanges_1.22.10        
[13] BiocGenerics_0.10.0    

loaded via a namespace (and not attached):
 [1] AnnotationDbi_1.26.0 Biobase_2.24.0       DBI_0.3.0           
 [4] RColorBrewer_1.0-5   RSQLite_0.11.4       XML_3.98-1.1        
 [7] XVector_0.4.0        annotate_1.42.1      genefilter_1.46.1   
[10] geneplotter_1.42.0   grid_3.1.0           lattice_0.20-29     
[13] locfit_1.5-9.1       stats4_3.1.0         survival_2.37-7     
[16] xtable_1.7-4 


 

deseq2 bugs • 1.4k views
ADD COMMENTlink modified 4.9 years ago by Michael Love25k • written 4.9 years ago by lmf.bill@gmail.com10

Can you please elaborate on your question? What does "report more differentially expressed genes than before" mean? Before what? Are you talking about using two different versions of DESeq2, or DESeq vs DESeq2? Please also provide details of your analysis, ie: experimental design and the code you are using to run your analysis. You also say that "actually, it is not reasonable" -- can you explain what you are seeing that is not reasonable? I mean: what is the thing in particular that tells you the result is not reasonable?

As it is now, your question is really too vague for anybody to be able to provide any meaningful help.

ADD REPLYlink written 4.9 years ago by Steve Lianoglou12k
Answer: DESeq2 more differentially expressed genes
0
gravatar for lmf.bill@gmail.com
4.9 years ago by
United States
lmf.bill@gmail.com10 wrote:

Thanks your interest my question.

I said DESeq2 reported more differentially expressed genes when I compared with DESeq.

My only concern is the DESeq2 output "padj column  is different from the values I am calculating using the p.adjust (pvalue column)". Thanks.

 

 

ADD COMMENTlink written 4.9 years ago by lmf.bill@gmail.com10

Just for the record, you never said "when I compared with DESeq", you just said "I used DESeq2 and found it report more deferentially expressed genes than before. Actually, it is not reasonable." Your "before" might have been an earlier version of DESeq2 or something else entirely ... it just wasn't clear.

I was also interesting in understanding how you made the call for the new DESeq2 result you were concerned about to be "not reasonable" -- perhaps you had some knowledge of the expected outcome of your experiment that was contradicting the results you obtained, or?

Also, to keep the boards "tidy/useful" these type of follow up comments should be added as comments/replies to my other comment, not as an answer to your question.

ADD REPLYlink modified 4.9 years ago • written 4.9 years ago by Steve Lianoglou12k
Answer: DESeq2 more differentially expressed genes
0
gravatar for Michael Love
4.9 years ago by
Michael Love25k
United States
Michael Love25k wrote:

DESeq2 is expected to be more sensitive than DESeq. This is described and demonstrated in our preprint referenced in our documentation: 

http://dx.doi.org/10.1101/002832

The padj column of the results object is different than p.adjust(res$pvalue) because the results() function uses independent filtering (by importing functionality from the genefilter package) to optimize the number of rejections at a given threshold on FDR. This is described in ?results and in the vignette Section 3.8 "Independent filtering of results".

ADD COMMENTlink written 4.9 years ago by Michael Love25k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 16.09
Traffic: 119 users visited in the last hour