Question

DEGs with uninsteresed sample files

0

Entering edit mode

3.4 years ago

jkim ▴ 190

Hi all,

I am working on some bulkRNAseq data from GEO and my sample description looks like this. I am interested in A vs B, A vs C and B vs C but I am not interested in sample D.

My question is whether I should include sample D for edgeR or DESeq2 to help them estimate some parameters such as commona variance or something. I have learnt about linear regression so I think I should include sample D(that's what the class taught me - reducing SE) but I am not 100% sure that it applies to DEGs.

enter image description here

design bulkRNAseq DEG matrix • 949 views

ADD COMMENT • link updated 3.4 years ago by swbarnes2 14k • written 3.4 years ago by jkim ▴ 190

score 2 · Accepted Answer · 2021-08-23

2

Entering edit mode

3.4 years ago

swbarnes2 14k

If the D samples don't have batch issues, and a PCA shows that they aren't way different from other others, or have way higher variability than the others, you should leave them in.

ADD COMMENT • link 3.4 years ago by swbarnes2 14k

0

Entering edit mode

Thank you for your suggestion. Yes, looking at a PCA plot will help for sure. Do you mind me asking a question? I would like to know what your reasoning about leaving them in is

ADD REPLY • link 3.4 years ago by jkim ▴ 190

1

Entering edit mode

Same as yours. See also: http://bioconductor.org/packages/release/bioc/vignettes/DESeq2/inst/doc/DESeq2.html#if-i-have-multiple-groups-should-i-run-all-together-or-split-into-pairs-of-groups

ADD REPLY • link 3.4 years ago by swbarnes2 14k