SNV

Owned by Cyrielle Kint
Last updated: Nov 24, 2021
4 min read

The automated pipeline firstly performs SNV variant filtering according to the criteria listed below. These initial filtering steps are followed by ranking of the remaining variants based on the combination of variant annotations, including info on predicted variant effect and phenotype overlap. A variant’s rank reflects the proposed relevance of the variant for the patient’s phenotype according to Moon.

The following criteria are taken into account for filtering SNVs:

Variant quality

  • Sequencing depth (DP in VCF) > 3 (if QUAL > 30) or > 8 (if QUAL =< 30)

  • Only variants with allele depth (AD in VCF) > 2 per allele (if QUAL > 30) or > 4 per allele (if QUAL =< 30) are retained. This criterium is not applied for mitochondrial variants.

  • Alternate allele ratio (AD of allele 1 divided by total AD) needs to be between 0.2 and 0.8. This criterium is not applied for mitochondrial variants.

  • genotype quality (GQ) >=20

  • Multiallelic sites (positions with at least 5 different Alt alleles based on samples uploaded on the server) are excluded, except for variants that are considered LP/P by Moon (see section Variant classification below).

If prefiltering is applied during variant calling, Moon will only consider variants with a PASS in the VCF Filter field for annotation and analysis. If you wish to analyse all variants in your VCF irrespective of the prefiltering results of the variant caller, a 'PASS' or '.' should be provided in the Filter field for each variant.

Low quality input VCF files might lead to false positive and false negative interpretation results. It is recommended to assess the overall quality of the input VCF prior to variant interpretation.

Population frequency

Common variants in the general population (>2% in gnomAD) are excluded except for a subset of more common variants with known pathogenic classification on ClinVar. This subset of common pathogenic variants includes all variants in the ClinGen exception list for ACMG BA1 (see https://www.clinicalgenome.org/working-groups/sequence-variant-interpretation/)

Additional variable frequency thresholds are applied depending on the inheritance pattern of the annotated disease.

Mitochondrial variants listed as polymorphism on Mitomap and having an allele frequency > 0.2% in GenBank are filtered out. In addition, mitochondrial variants for which gnomAD contains homoplasmic observations are also filtered out except if known pathogenic. For variants with reported disease associations (either ‘reported’ on Mitomap or classified LP/P on ClinVar), a 2% Mitomap threshold is applied.

Variant classification

Variants with either Benign or Likely Benign classifications on ClinVar, Invitae KB or within the lab’s own knowledge base, are excluded from the analysis. Pathogenic or Likely Pathogenic variants on ClinVar, Invitae KB or the lab’s knowledge base are taken into account to optimise both filtering and ranking. For variants with conflicting interpretations, the classification is prioritised based on: Lab’s internal knowledge base > Invitae KB > ClinVar. In the case of variants with truly conflicting interpretations within ClinVar (at least 1 B/LB combined with at least 1 P/LP classification), the most severe classification (LP/P) is considered by Moon for the analysis.

Variant effect

Variants in protein or RNA coding regions, variants predicted to have a significant effect on splicing (located either within or outside of the splice region) and known pathogenic variants in non-coding regions are retained. For mitochondrial variants, all variants located within non-protein coding genes are retained if being in accordance with other filter criteria. De novo splice region variants are also retained even if splice predictions are not significant, if being in accordance with other filter criteria.

Inheritance

Only variants for which the zygosity of the called variant fits the inheritance pattern of the annotated disease are retained in the ‘SNV’ shortlist, with addition of (potentially) compound heterozygous variants.

Heterozygous variants in genes associated with recessive disorders are displayed in the ‘Carrier’ list, if in accordance with other filter criteria.

Familial segregation

In family analyses, co-segregation of the variant with the phenotype (ie. healthy or affected status of included family members), according to autosomal dominant, autosomal recessive, X-linked dominant, X-linked recessive or mitochondrial inheritance patterns, is taken into account. Segregation is not only applied as a strict filter criterion, thereby also ensuring that causal mutations following non-Mendelian inheritance (eg. with incomplete penetrance) can be identified in family analyses.

Confirmation of parentage is needed for interpretation of results of family analyses. Misattributed parentage can impact the results.

Phenotype overlap

In a final filter step, the phenotype overlap is scored between the input HPO terms describing the patient’s phenotype and known disease manifestations of the annotated disorder. Less stringent phenotype filtering is applied for variants that are either classified as LP/P, or that cause loss of function (LOF) of a gene, for which the annotated disorder is known to be caused by a LOF pathogenic mechanism.

Variants in genes for which the phenotype match with the annotated disease is considered too limited by Moon, are excluded.