Support for bcftools call -m (multiallelic caller) in bamCaller.py

[FWFKI]

Hello, and thank you for providing this useful tool.

I attempted to run the bamCaller.py pipeline using the multiallelic caller method in bcftools call (i.e., using the -m option). However, I found that no sites were output in either the resulting VCF file or the generated mask file.

Upon checking, it seems that this issue occurs because the INFO/FQ tag, which is used by bamCaller.py for filtering, is not produced by the multiallelic caller in bcftools.

If possible, I would prefer to use the multiallelic calling method, as it is the newer　approach. Would it be possible to implement compatibility with this method in bamCaller.py, or could you suggest a workaround to enable its use?

Thank you very much for your time and support.

Best regards,
Keita Ido

[stschiff]

Hi. No, I can't help with implementing this. I took the entire filtering procedure from Li and Durbin 2011, and the consensus quality approach using the FQ tag is essential for that. I wouldn't know what an equivalent with the multi allelic caller would be. I also think that for single-sample calling the -c caller is actually more appropriate.

[FWFKI]

Thank you very much for your response and clarification. I understand the background and appreciate your time. I also agree that the -c caller makes more sense for single-sample calling. When preparing the VCF for MSMC, I carried out joint calling on multiple samples using the -m caller, so I wanted to apply the same settings here for consistency.