We performed NGS using cells collected from mice in a xenotransplantation study.

As such, the FASTQ files contain reads of DNA from both mice and human cells.

I expect ~30% of reads are contaminated by mice cells.


Is it wrong to directly align using the indexed human reference with bowtie2?

Is there a way to use two indexes at once with bowtie2? The goal here would be to get the % of reads that map to each genome.


1 Answer 1


I am not aware of a method using two indices in bowtie2 but here is a simple workaround:

Get human reference genome as fasta and suffix all fasta names with _human. Do the same with the mouse genome using _mouse. cat both together and build an index. Then you can later track back whether the alignment was done to human or mouse.

Edit: Be aware though that for those parts of the genomes that have high sequence similarity you will get multimapping reads with MAPQ of 0 (or at least very low).

  • 1
    $\begingroup$ This is generally the right way to go about it. Different methods will handle problems like sequence similarity or comparing mapping qualities differently, so Nomad420 you should be aware of what different tools do. One place to start looking is the Xenome tool. It is a command line tool that does what ATpoint suggests and classifies reads based on how they map to each reference genome. $\endgroup$ Apr 27, 2022 at 16:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.