error of the BAM files using MetaBAT2

[Holly0929]

Hi, I am currently using MetaBAT2 for binning genomes from my assembly.fa and sorted BAM files. However, each time I run the command (" jgi_summarize_bam_contig_depths --outputDepth metabatdepth.txt .bam " or "runMetaBat.sh assembly.fa .bam" ), some error message keeps popping up like this:

Warning: SAM 1.3 M not 1.4 =/X and there is no NM aux field. PercentID will be invalid!!! SN1040:128:HY35KBCXY:2:2106:4378:93142_2:N:0:GGACTCCT+TATCCTCT^C

May I know what was wrong? and how to solve this problem?

Thank you.

[Rob Egan]

The issue is that the SAM file does not have enough information to determine the % identity that the read has to the reference and metabat uses this to ignore reads that map, but map poorly, primarily because the read is from a closely related species which happens to be missing from the assembly.

You can either: 1) map with an aligner that supports the SAM format >= 1.4 (which uses = or X in the cigar string for match and mismatch respectively (instead of M from previous versions of the SAM format which is ambiguous)). bbmap is a my recommendation here.

2) execute samtools calmd on your SAM file which uses the reference to include the auxiliary fields NM and MD which will allow metabat's tool to calculate the % ID.

-Rob

[Holly0929]

Thanks a lot for your quick response, Rob Egan. I will try another mapping tool and keep updated with you.