Bowtie2 which reference is best ?

Question

[Aki]

Hello

I am trying to learn Bowtie2. When I compared the overall alignment rate by bowtie2, there is a significant difference between the result of GRCh37 index and GRCh38 index. The overall alignment rate to GRCh37 is 98%, but that to GRCh38 is 18%. Why does this happen? The author who used this chip-seq data adopt GRCh37. Does anyone have the answer?

Answer 1

[ATpoint]

This must be something else, these genomes are too similar for such dramatic results. Be sure that you did not accidently used GRCm38 the mouse genome. Please check these obvious things first, like listing all chromosomes of the fasta files, that will show if one is mouse. running and posting the output of samtools idxstats for both bam files.

Answer 2

[colindaven]

It's might be the ALT contigs https://gatk.broadinstitute.org/hc/en-us/articles/360035890951-Human-genome-reference-builds-GRCh38-or-hg38-b37-hg19

Though the mouse genome is also a possibility as ATpoint says.

To map chip-seq exclude the ALT contigs. I don't think bowtie2 is an ALT aware mapper.

Also play with the settings for bowtie2 - you can allow multiple mappings if you like.

[ATpoint]

With ALT the reads coming from those regions would still be mapped but multimapping between ALT and non-ALT, no? Multimappers count as mapped afaik.