Question

Log2 foldchange in DESeq2

0

Entering edit mode

Sushant Pawar • 0

@sushant-pawar-9287

Last seen 7.3 years ago

Nashik

Hello all,

I want to ask a question about Log2FoldChange in DESeq2. In many cases for the similar counts of Case and Control if the count shows down regulation but the DESeq2 give the output Up regulation, it gives us trouble for further analysis, so can anybody tell me how to deal with this problem.

here is some sample of our data.

GENES	logfc	adjpv	Counts_Control	Counts_Case
xyz1	2.000605	0.000001	2231	1330
xyz2	2.001338	0.000001	290	173
xyz3	2.012326	0.000001	92	55
xyz4	2.005319	0.000001	433	259
xyz5	2.012833	0.000001	75	45
xyz6	2.006388	0.000001	30	18
xyz7	-2.34172	0.000001	2111	62
xyz8	-2.34066	0.000001	1225	36
xyz9	-2.33344	0.000001	238	7
xyz10	-2.33951	0.000001	1768	52

as you can see in above table xyz1-xyz6 results are Down Regulated but the DESeq2 give the Up Regulation.

where xyz7-xyz10 shows the correct result. Please help me in this problem.

Waiting for positive response.

deseq2 R log2fc • 2.3k views

ADD COMMENT • link updated 8.4 years ago by Michael Love 41k • written 8.4 years ago by Sushant Pawar • 0

score 1 · Answer 1 · 2015-11-30

1

Entering edit mode

Michael Love 41k

@mikelove

Last seen 4 hours ago

United States

You haven't really provided enough information here for someone to help. You should provide all your code when posting to Bioc support. And additionally it's helpful to see sessionInfo().

If I were to guess based on what I can see here, I'd point out that the log2 fold change is based on normalized counts, not raw counts. Raw counts do not account for sequencing depth.

ADD COMMENT • link 8.4 years ago Michael Love 41k

0

Entering edit mode

Thank you for reply,

====================================================================

library(DESeq2)
directory<-"/home/projects_110_AmpliseqRNA/all_samples_with-replicates/LN_R2"
sampleFiles <- grep("treated",list.files(directory),value=TRUE)
sampleCondition<-c("treated","treated","treated","untreated","untreated","untreated")
sampleTable<-data.frame(sampleName=sampleFiles, fileName=sampleFiles, condition=sampleCondition)
sampleTable
ddsHTSeq<-DESeqDataSetFromHTSeqCount(sampleTable=sampleTable, directory=directory, design=~condition)
colData(ddsHTSeq)$condition<-factor(colData(ddsHTSeq)$condition, levels=c("untreated","treated"))
dds<-DESeq(ddsHTSeq,fitType="mean")
res<-results(dds)
res<-res[order(res$padj),]
head(res)

mcols(res,use.names=TRUE)
write.csv(as.data.frame(res),file="Raw_File_deseq2.csv")

FileforIpathway=res[,c(2,6)]
write.csv(as.data.frame(FileforIpathway),file="File_deseq2.csv")

pdf("DEseq2_MA_plot.pdf")
plotMA(dds,ylim=c(-10,10),main="DESeq2")
dev.off()

pdf("DESeq2_Dispersion_plot.pdf")
plotDispEsts(dds)
dev.off()

====================================================================

This is the code which i am using for analysis.

ADD REPLY • link 8.4 years ago Sushant Pawar • 0

0

Entering edit mode

Take a look at these genes using plotCounts (see vignette). I think you'll see that the LFC accords with the normalized counts across the different samples.

ADD REPLY • link 8.4 years ago Michael Love 41k