Allele option doesn't restrict input to actual human class I alleles

[susannasiebert]

As far as I understand it from your paper, BigMHC supports predictions for all human class I alleles. However, it doesn't restrict the input provided to the -a option to only human class I HLA alleles. I was able to provide non-human alleles, human class II alleles, as well as nonsense words and a prediction result was returned. Is this intended or should this option be limited to only inputs that are actual class I human HLA alleles?

[benjaminalbert]

To handle different MHC naming schemes and possibly MHCs that are not found in our MSA, we have a fuzzy string matching method called nearestMHC in the src/mhenc.py file. However, as you note, this fuzzy matching protocol does not filter out completely invalid inputs.

[RachelKarchin]

@susannasiebert Is this issue an obstacle to integrate BigMHC into pVacTools? If yes, we will come up with a workaround for you. Please let us know.

[susannasiebert]

For all algorithms pVACtools currently support we maintain files for supported alleles for each algorithm in order to fail early when someone supplies a nonsense value for the allele or skip prediction calls for unsupported alleles. I think for BigMHC I will just manually create such a file with all known human class I alleles.

However, I do think it would be worthwhile to have BigMHC fail if a user supplies non-human or class II alleles since right now predictions will be made and that might be confusing to users. Unless I misunderstood the paper and such predictions are valid?

[benjaminalbert]

The list of alleles used in the MSA can be found here: https://github.com/KarchinLab/bigmhc/blob/master/data/pseudoseqs.csv

I've add some text to the README describing the fuzzy string matching.

[RachelKarchin]

One caveat to that - the alleles used in the MSA include many that are not human.

[susannasiebert]

One caveat to that - the alleles used in the MSA include many that are not human.

Can BigMHC be used with non-human data/alleles then?

[benjaminalbert]

One caveat to that - the alleles used in the MSA include many that are not human.

Can BigMHC be used with non-human data/alleles then?

Although BigMHC can consume non-human data, we have not trained or validated on non-human data, so I would not recommend it.

[Stikus]

@susannasiebert As far as I see - your fork https://github.com/griffithlab/bigmhc is 10 commits ahead of, 4 commits behind.

Also, I see opened PR #8

What is the current state of this issue and PR? Should we use your fork or latest bigmhc ?