Supplied depth.txt file vs. supplied bam files

[Former user]

Hello,

I noticed that there's a difference in the number of bins when I supply a depth.txt file to metabat vs when I supply sorted and indexed bam files, even if the depth.txt file was generated using jgi_summarize_bam_contig_depths. Can you please tell me why?

Additionally, are there any fixes for having bam files with incorrect refpos and len issues?

Thanks!

[Rob Egan]

There is a random component to the binning stage. If you set the --seed argument in metabat2 to something non-zero, then the bins should be identical given the same input files.

You should also verify that the depth.txt is the same one which is generated when you give the runMetaBat wrapper script the bam files. It is possible that you used different options to generate them.

samtools calmd can sometimes fix improper BAM files from imperfect aligners, but really it is the fault of the aligner, so notify the developer of that tool.