search

EnvGen / POGENOM

Population genomics from metagenomes
BSD 2-Clause "Simplified" License
10 stars 4 forks source link

add bwa-mem2 support to improve mapping speed #5

Closed jianshu93 closed 2 years ago

jianshu93 commented 3 years ago

Dear author,

It would be nice to add bwa-mem2 support for linux (conda install -c bioconda bwa-mem2) for large dataset like a few hundred samples. Bowtie2 does not scale well with increasing threads (core dumped for large threads). Should be very easy, just the index and mapping step. All others are the same:

bwa-mem2 index ${output}/all_mags_rename.fasta bwa-mem2 mem -t $threads -p ${output}/all_mags_rename.fasta $reads1 $reads2 > ${output}/all_mags_rename.sam

Thanks,

Jianshu

jianshu93 commented 3 years ago

bam filtering can be then done using coverm filter acording to alignment ratio and identity.

lfdelzam commented 2 years ago

Thanks. So far bowtie2 has worked very well with big datasets. In addition, Bowtie2 allows for setting the flag –score-min (it gives the minimum score that is allowed to report an alignment) to a user-defined identity threshold (e.g 95%).

lfdelzam commented 2 years ago

the new version 0.8.4 includes BWA-MEM2 + CoverM