Questions and Improvement Requests for BamFilter

[BenjaminGuinet]

MAN-2-1_filterBAM.txt

Dataset.zip

Dear BamFilter developers,

We are a group of researchers who have been using your tool in our lab and have found it very useful. Thank you for providing such a helpful resource.

While working with BamFilter, we encountered a few questions and potential areas for improvement:

1. Empty result columns
   In some cases, the following columns are empty in the output: read_aln_score, edit_distances, read_ani_mean, read_ani_std, read_ani_median (see attached table named MAN-2-1_filterBAM.txt in the zip folder).

2. High memory usage
   The memory requirements can sometimes be quite high (several GB of RAM). Is there a way to reduce memory usage, perhaps by using the low-memory option we noticed in the code?

3. Header-related error
   When BAM headers include contigs from previous reference genomes with no mapped reads, we encounter the following error:

pandas.errors.IntCastingNaNError: Cannot convert non-finite values (NA or inf) to integer

Could it be possible to handle this case ? We would like to retain contig names in the header without causing an error.

Command used:
filterBAM filter --bam ${bam} --stats ${prefix}_filterBAM.txt

Files referenced:

• MAN-2-1_filterBAM.txt: table with missing columns, generated from MAN-2-1.NC_001640.dedup.bam.

• MAN-2-1.NC_001640_com.dedup.bam: BAM file that triggers the header-related error.

Thank you very much for your time and for maintaining this tool. We look forward to your feedback.

Best regards,

Benjamin