Is it possible to resolve multi-nucleotide variants (MNV) to biallelic records?

[rajwanir]

Hi @freeseek

I notice that some MNV variants are often assayed using Infinium arrays. These MNV records are represented with a single record in manifest with a single allele but gets translated into multi-allelic in the gtc2vcf. My understanding is that a single probe can only interogate a single allele. May I please seek your help in understanding why and how these MNV records gets translated into multi-allelics?

For example:

CSV manifest record:

IlmnID	Name	IlmnStrand	SNP	AddressA_ID	AlleleA_ProbeSeq	AddressB_ID	AlleleB_ProbeSeq	GenomeBuild	Chr	MapInfo	Ploidy	Species	Source	SourceVersion	SourceStrand	SourceSeq	TopGenomicSeq	BeadSetID	Exp_Clusters	Intensity_Only	RefStrand
5:112838101_MNV-0_B_R_2716756713	5:112838101_MNV	BOT	[T/C]	37787959	CTCTCCAAACTTCTATCTTTTTCAGAACGAGAACTATCTAAGCTTCCTCT			38	5	112838101	diploid	Homo sapiens	clinvar	0	BOT	TCCGCGTTCTCTCTCCAAACTTCTATCTTTTTCAGAACGAGAACTATCTAAGCTTCCTCT[T/C]NNNAGGAGCTGGGTAACACTGTAGTATTCAAATATGGTGAAAGGACAGTCATGTTGCCAG	CTGGCAACATGACTGTCCTTTCACCATATTTGAATACTACAGTGTTACCCAGCTCCTNNN[A/G]AGAGGAAGCTTAGATAGTTCTCGTTCTGAAAAAGATAGAAGTTTGGAGAGAGAACGCGGA	1984	3	0	-

VCF record:

#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO
chr5	112838101	5:112838101_MNV	C	A,G	.	.	GC=0.4125;ALLELE_A=1;ALLELE_B=2;FRAC_A=0.310924;FRAC_C=0.193277;FRAC_G=0.235294;FRAC_T=0.260504;NORM_ID=9;BEADSET_ID=1984;ASSAY_TYPE=0

From the chip: ftp://[email protected]/Public_Docs/Genotyping_Array_Support_Files/Global%20Screening%20Array/Global%20Screening%20Array%20v3/GSAv3%2BConfluence/NCI_custom_booster_20032937X371431_A1.csv

The behaviour of Illumina Dragen array gtc-to-vcf is identical (i.e. it also outputs multi-allelic variant in this scenerio).

Thank you.