removing outliers before running DESeq2

Question

[nazaninhoseinkhan]

Dear all,

After running DESeq2 program successfully I found out that I had to detect and remove outliers from my htseq count datasets before running DESeq2. However, I am a little confused about how to perform this task.

Can u advise me the most straightforward way to do this?

Thank u in advance

Nazanin Hosseinkhan

[firatuyulur]

Why do you think you need to remove outliers? I tend to remove the genes that doesnt have more than 5 counts on average across all samples but nothing more.

[nazaninhoseinkhan]

Thank u so much. Yes, I've already removed genes with lower than 10 reads

[Arindam Ghosh]

Why 10? Why not any other?

Answer 1

[Arup Ghosh]

There is no need to remove out layers but it's better to remove the technical artifacts(noise incorporated by the sequencer). Genes/transcripts having read count within 0-10 rage is generally considered as artifacts(if the expression pattern is inconstant throughout all the samples). For more details check the following thread.

https://support.bioconductor.org/p/95755/

[nazaninhoseinkhan]

Thanks a lot for your advice.