Tim Dunn

Adding the option `--engine vcfeval` usually improves detection of trickier variants. If there are any partially overlapping calls, this might be able to detect them using a combination of this...

Sounds good, thanks for the quick response! I'll start contributing after I finish work on another project.

Okay, thanks for responding so quickly! You mention that "somatic calling isn't supported very well" because setting the ploidy to 4 is kind of a hacky fix to allow Nanopolish...

Any updates on the current status of the Haar wavelets? I would like to use it, but can't find in the docs if it's been added yet. Thanks!

I'm running into the same errors. Thanks for suggesting the easy fix

Yes, I am currently working on on extending vcfdist to work with unphased VCFs. Doing so requires significant changes to the underlying algorithm, and I will be working on this...

For your evaluations, will you be able to obtain a phased ground truth VCF? Or would both your query and truth VCFs be unphased? That would make things considerably harder.

This is my first time writing Rust, so please feel free to make suggestions for better style/design/etc. A note on the new tests: Since `Sample`s do not allow empty barcodes:...

Is the plan for this issue to remove `typing.List` (and other types) from `inspect._get_parser` only, or to also more generally replace all usage of `typing.List` with `list`?

Leafcutter's `prepare_results.R` script outputs an RData file that contains a matrix `counts` of per-junction coverage counts. Using this matrix and [`countdata`](https://cran.r-project.org/web/packages/countdata/index.html)'s `bb.test` module, we were able to obtain p-values for...