Hi,

In the differential expression analysis we are interested in studying the variation of gene expression between males and females in mouse brain. We ran DESeq2 genome-wide while filtering out results for keeping only sexual chromosomes in order to highlight genes with significant variations.

However, after investigating the results we have observed weird results with two genes on the Y chromosome having significant LFC between males (reference category) and females (comparison category), which is little bit unexpected from my taste. Indeed, I am expecting to have only genes on the X chromosome not on the Y, when comparing males to females.

Are there any explication for this result ? Is it a design issue ? It is the first time I work on sexual chromosomes and I probably miss something important here.

I can provide additional material if needed,
Loic

You'd have to say which genes. There are a set of genes that are in the pseudoautosomal region (PAR) that could easily be differentially expressed. As an example, AKAP17A:

> library(TxDb.Hsapiens.UCSC.hg38.refGene)
> library(org.Hs.eg.db)
> select(org.Hs.eg.db, "AKAP17A", "ENTREZID","ALIAS")
'select()' returned 1:1 mapping
between keys and columns
    ALIAS ENTREZID
1 AKAP17A     8227
> gns <- genes(TxDb.Hsapiens.UCSC.hg38.refGene, single.strand.genes.only = FALSE)
> gns["8227"]
GRangesList object of length 1:
$`8227`
GRanges object with 2 ranges and 0 metadata columns:
      seqnames          ranges
         <Rle>       <IRanges>
  [1]     chrX 1591604-1602518
  [2]     chrY 1591604-1602518
      strand
       <Rle>
  [1]      +
  [2]      +
  -------
  seqinfo: 711 sequences (1 circular) from hg38 genome

Unexpected results may also raise from the assignment of a wrong biological sex to the samples. This can be easily detected in transcriptomics data by looking up whether the expression of sex-specific genes matches the given biological sex of the sample.