Updated Cyvcf2 Adds MNPs (Multiple Nucleotide Polymorphism) -Will Break Tests

Somewhere between our locked version of Cyvcf2 (<0.10.0) and the latest version, support for MNPs was added. Displayed below is the result of basic content search.

Newest

cyvcf2% grep -c mnp *
__init__.py:0
__main__.py:0
grep: __pycache__: Is a directory
cli.py:0
cyvcf2.c:36
cyvcf2.cpython-39-darwin.so:3
cyvcf2.pxd:0
cyvcf2.pyx:3
helpers.c:0
helpers.h:0
relatedness.h:0
grep: tests: Is a directory

<0.10.0

cyvcf2% grep -c mnp *
__init__.py:0
__main__.py:0
grep: __pycache__: Is a directory
cli.py:0
cyvcf2.c:0
cyvcf2.cpython-38-darwin.so:0
cyvcf2.pxd:0
cyvcf2.pyx:0
helpers.c:0
helpers.h:0
relatedness.h:0
grep: tests: Is a directory

Running the test suite will eventually break on test tests/adapter/mongo/test_query.py::test_query_snvs_by_coordinates when calling the constant CALLERS with argument mnp. This because variant.var_type which previous was set to SNP or SNV will now may be set to MNP. I have come up with two simple solutions, but unsure which one is correct or best approach.

A. Add Support for MNP

Add support for mnp:s in CALLERS constant. /constants/__init__.py. –I am unsure how this should look.

CALLERS = {
    "snv": [
        {"id": "gatk", "name": "GATK"},
        {"id": "freebayes", "name": "Freebayes"},
        {"id": "samtools", "name": "SAMtools"},
        {"id": "bcftools", "name": "Bcftools"},
        {"id": "deepvariant", "name": "DeepVariant"},
    ],
    
    "cancer": [
        {"id": "mutect", "name": "MuTect"},
        {"id": "pindel", "name": "Pindel"},
        {"id": "gatk", "name": "GATK"},
        {"id": "freebayes", "name": "Freebayes"},
    ],
    "cancer_sv": [
        {"id": "manta", "name": "Manta"},
        {"id": "gatk", "name": "GATK"},
    ],
    "sv": [
        {"id": "gatk", "name": "GATK"},
        {"id": "cnvnator", "name": "CNVnator"},
        {"id": "delly", "name": "Delly"},
        {"id": "tiddit", "name": "TIDDIT"},
        {"id": "manta", "name": "Manta"},
    ],
    "str": [{"id": "expansionhunter", "name": "ExpansionHunter"}],
}

B. Cast MNP:s to SNPs

To my understanding a MNP is a set of SNVs. Can they be treated as equivalent? In parser/variant/variant.py add the following:

    if not category:
        category = variant.var_type
        if category == "indel":
            category = "snv"
        if category == "snp":
            category = "snv"
        if category == "mnp":
            category = "snv"

[mikaell]

It will be a bug in the moment we de-freeze the CyVCF2 lib, it’s not a bug at the moment, right @mikaell ?

No, not at the moment.

[northwestwitch]

This one was fixed by #3114. Closing!