I´m trying to align paired-end reads (illumina) coming from a metagenomics experiment to a custom database of bacterial species.
I´m using bwa-mem (althought could change program if required) to align reads to my genomes. It works pretty well but i have some reads mapping to two different bacterial species. I know this is normal and occurs because those genomes a very close related.
In my case i have Species-A mapped by 2200 reads where as close related speciesB is mapped by 800reads. Only speciesA is present in my sample so clearly speciesB is a false positive.
Is there any way to try to reduce false positives by tuning bwa in order to minimize the impact of cross mapping reads ??
thanks , david