Error with haplotagged BAM files - Unexpected HP tag format

[Audald]

I am facing a problem when running aligned_bam_to_cpg_scores for haplotagged BAM files. This is the command used:

/path/pb-CpG-tools-v2.3.0-x86_64-unknown-linux-gnu/bin/aligned_bam_to_cpg_scores --bam /path/sample.haplotagged.bam --output-prefix /path/phased/BEDs/sample.pbmm2 --model /path/pb-CpG-tools-v2.3.0-x86_64-unknown-linux-gnu/models/pileup_calling_model.v1.tflite --threads=8

And the error is as follows:

thread '<unnamed>' panicked at 'Unexpected HP tag format in read ABCD I32(2)', src/meth_read_processor.rs:1128:9
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

The HP tag in read ABCD is HP:i:2, which seems aligned to what is supposed (see #43). We then realised that some of the reads are not HP-tagged (ambiguous reads?) and thought that could be problematic. However, the problem is still there when the untagged reads are removed from the BAM file.

The phasing was performed with HiPhase.

[ctsa]

Thanks @Audald - there is an issue here handling the binary representation of the integer in the HP field -- it looks like whatshap always writes this as a uint8, and here it's an int32. I'll fix the method to handle all int representations.

[ctsa]

Here's the update with fixed HP parser, hopefully this should take care of the issue:

https://github.com/PacificBiosciences/pb-CpG-tools/releases/tag/v2.3.1

[Audald]

Hi @ctsa, thanks for the prompt fix. It does work!

[ctsa]

Good to hear. Closing as resolved.