search

jtleek / sva-devel

28 stars 45 forks source link

Combat returns negative values. #52

Closed lincj1994 closed 2 years ago

lincj1994 commented 2 years ago

Hi. I have performed the batch effect correction using combat but it returned negative values for some genes in specific samples. I'm wondering how to deal with it? Replace the negative values with zero OR with the original values before combat? Btw：

  1. Should I log2 transformed the input data (FPKM) before the combat function? OR it doesn't matter whether to log-transform the data or not.
  2. If the input data was log transformed (log FPKM), can I exp the data after combat to get a combated FPKM (exp(log2FPKM))? Thanks.
KC-Lan commented 2 years ago

Hi, For your first question, maybe you can check this: Combat generating negative values, should I log2 transform first?

wevanjohnson commented 2 years ago

If you are working with RNA-seq data, check out ComBat-Seq:

https://academic.oup.com/nargab/article/2/3/lqaa078/5909519

On Mar 29, 2022, at 1:11 AM, KC-Lan @.***> wrote:

Hi, For your first question, maybe you can check this: Combat generating negative values, should I log2 transform first? https://support.bioconductor.org/p/73266/ — Reply to this email directly, view it on GitHub https://github.com/jtleek/sva-devel/issues/52#issuecomment-1081561603, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACMBWPAMT2R2G3QBUPZOD6TVCK3JRANCNFSM5QGPKUAQ. You are receiving this because you are subscribed to this thread.

lincj1994 commented 2 years ago

Thanks for your reply!