True deletion discarded in final vcf

[yiqijiang17]

I am using SvABA for structural variant calling on a bacterial genome with gene edits. I have noticed that all edited regions are being discarded in the variant calling, even though the edits appear successful based on mapping depth.

Specifically, I deleted gene A at positions pos1-pos2. In the bam file, I see mapping depth of 0 across this region, indicating a successful edit. I can also see the deleted region assembled as a contig in the alignments.txt.gz file, named "Global BP: chr:pos1(+) to chr:pos2(-) c_1_343001_368001". This contig has ~60X depth.

However, I cannot find this deletion in the final bps.txt.gz and unfiltered sv.vcf files. The log file also does not mention why this contig was discarded:

...assembled 1 contigs for c_1_343001_368001
...assembled 8 contigs for c_other

Do you have any suggestions on why this candidate variant was filtered out? And which SvABA parameters I could adjust to retain it in the final callset? I want to ensure I am capturing all real edits.

Thanks!