ComBat-Seq [Zhang2020]_ follows on the steps of ComBat, but targets specifically
RNA-Seq data. Conceptually, ComBat-Seq is based on the same mathematical
framework as ComBat, except that its replaces the normal distribution of
microarray data by a negative binomial distribution to account for the
specificities of RNA-Seq expression data. :func:`pycombat_seq` is a direct port
of ComBat-Seq to Python. Since ComBat-Seq relies on the Bioconductor
edgeR
package, the relevant parts of edgeR
have been ported
along. Closely following the original implementation in R, :func:`pycombat_seq`
has results very similar to those of ComBat-Seq in terms of batch effects
correction. Additionally, :func:`pycombat_seq` is as fast, if not faster, than
the original implementation in R. It also features additional capabilities, such
as fixing a given batch as reference.