I know this topic has been touched in the past but I couldn't find an answer to my question. Background: in my lab, we are thinking to move from chip-seq of histone marks to atac-seq. We have data on CD4+ T cells and we would like to compare what we found with an already published dataset in the same cell type (Buenrostro et al. 2013: GEO: GSE47753). In addition, I would use this as a test to see if we have the capabilities of doing an atac-seq analysis.
So I downloaded the 6 sra files regarding the CD4+ T cells, produced the fastq files with relative quality control and aligned to hg19 using bowtie2 (
-X2000 ). I know in the paper they use bowtie instead of bowtie2 but I didn't think that could make a big difference at this step.
Looking at the SAM fils after the alignment, I realized that more than 90% of the reads are mapped to mitochondrial DNA leaving me with 50 to 200 thousand reads to do peak calling.
I tried to align using bwa mem but no much difference was found.
In the paper, they declare to have an average around 46% of reads mapping to mtDNA.
So the questions: what am I doing wrong? Did anyone else find the same thing? Am I misreading the methods of the paper?
Thanks in advance and I'm sorry for such a beginner question.