zhangyuqing / ComBat-seq

Batch effect adjustment based on negative binomial regression for RNA sequencing count data
154 stars 39 forks source link

Batch effect present for some biotypes after running CombatSeq #41

Open kvn95ss opened 5 months ago

kvn95ss commented 5 months ago

I'm working with smallRNA data and performed batch control with Combat_Seq (bioconductor version).

I used the below command -

library('sva')
corrected_batch <- ComBat_seq(as.matrix(counts), batch = coldata$Small_Seq_Batch)
corrected_batch <- as.data.frame(corrected_batch)

Here's a MDS plot before correcting - MDS_before_correction

And here's the after MDS_after_correction

What could cause this? I tried separating the biotypes first then performing the batch correction, but the outcome was pretty much the same.