extracting differential expressed genes from limma

Question

[zizigolu]

hi,

this is limma output of a comparison between treatment and control. i was going to extract genes with logFC > 2 but

head(data[,1:6])

          logFC   AveExpr        t      P.Value    adj.P.Val        B

260978_at 6.990868 10.370103 19.73560 2.029415e-08 0.0002399517 5.898061 255811_at 9.109741 8.188893 19.65074 2.103917e-08 0.0002399517 5.888268 250296_at 5.854323 10.503117 16.89666 7.411089e-08 0.0005634898 5.505260 253884_at 8.375813 7.832067 13.73106 4.109986e-07 0.0023437193 4.843450 247691_at 6.299296 10.104694 10.17111 4.648161e-06 0.0169450142 3.597549 259481_at -1.165086 7.677000 -9.87275 5.885539e-06 0.0169450142 3.456404

data=data[abs(data$logFC)>2,]

Error in mycounts$logFC : $ operator is invalid for atomic vectors

then how I can extract genes with logFC > 2 ???

thank you

Answer 1

[andrew.j.skelton73]

Use the topTable function from limma, and add in a fold change cutoff. Also, if the fold change is in log2, then looking for things greater than 2, means you're looking for 4 fold differences.

[zizigolu]

thank you, how i could know fold change is log2 or not?

[mastal511]

In the data you showed above, the first column is labelled logFC, in limma that is log2(fold change).

[zizigolu]

sorry, then if I select logFC > 1 that means I selected genes with two times fold change between treatment and control???

[mastal511]

Yes, that's right.