VariantColumnMapper

TODO -- update

ColumnMapper for columns that contain information about the variant(s) found in an individual. There are two main ways of using this class. If there is a column that contains an HGVS expression for the variant (without the transcript), then the transcript is indicated by the default_transcript argument.

HGVS vs Structural

pyphetools uses two different classes to ingest small variants (which must be encoded using valid HGVS notation or structural variants (which are not validated).

HGVS

The following code can be adapted to read HGVS-encoded variants. It assumes that there is a column in the input table called Variant. It first extracts a list of unique variants, and used variant validator to check them and to retrieve additional data such as the chromosomal position. The resulting Variant object is placed in a dictionary together with the original string used in the table (if necessary, one can correct minor errors such as the failure to use the "c." below).

Calling Variant Validator for HGVS-encoded variants

hg38 = 'hg38'
default_genotype = 'heterozygous'
WFS1_transcript='NM_006005.3'
vvalidator = VariantValidator(genome_build=hg38, transcript=WFS1_transcript)
variant_list = df['Variant'].unique()
print(variant_list)
variant_d = {}
for v in variant_list:
    if v == "1380del9":
        hgvs = "c.1385_1393del"
    else:
        hgvs = f"c.{v}"
    var = vvalidator.encode_hgvs(hgvs)
    variant_d[v] = var

For structural variants, pyphetools enocdes the variant using corresponding classes from Sequence Ontology with the StructuralVariant class, which has the static functions

• chromosomal_deletion
• chromosomal_duplication
• chromosomal_inversion

as well as a constructor that can take any relevant sequence ontology class.

chromosomal deletion

sv = StructuralVariant.chromosomal_deletion(cell_contents="46,XY.ish del(7)(p14.1)(RP11-816F16-)", 
````````````````````````````````````````````gene_id="HGNC:4319", 
                                            gene_symbol="GLI3")

The cell contents argument contains the variant name as used in the original table.

It is possible to mix HGVS and structural variants in Python code. An example follows.

HGVS and structural variants

struct_variants = { "rsa7p14.1(kit P179)x1",
                    "46,XY.ish del(7)(p14.1)(RP11-816F16-)",
                    "46,XX.ish del(7)(p14.1p14.1)(GLI3-)" }
gli3_symbol = "GLI3"
gli3_id = "HGNC:4319"
gli3_variants = df1['cDNA alteration'].unique()
gli3_variant_d = {}
for gli3v in gli3_variants:
    if gli3v in struct_variants:
        sv = StructuralVariant.chromosomal_deletion(cell_contents=gli3v, gene_id=gli3_id, gene_symbol=gli3_symbol)
        print(gli3v)
        gli3_variant_d[gli3v] = sv
    else:
        v = hgvsMapper.encode_hgvs(gli3v)
        gli3_variant_d[gli3v] = v

VariantColumnMapper

One the dictionary of variants has been constructed as above, we create the variant mapper as follows.

VariantColumnMapper constructor

variantMapper = VariantColumnMapper(variant_d=gli3_variant_d,
                                    variant_column_name="cDNA alteration",
                                    default_genotype='heterozygous'
                                   )                      

The VariantColumnMapper is now ready to go to map HGVS expressions in the column "cDNA alteration".