VEP-G2P: A Tool for Efficient, Flexible and Scalable Diagnostic Filtering of Genomic Variants

Anja Thormann, Mihail Halachev, William McLaren, David J Moore, Victoria Svinti, Archie Campbell, Shona M Kerr, Sarah Hunt, Malcolm G Dunlop, Matthew E Hurles, Caroline F Wright, Helen V Firth, Fiona Cunningham, David R FitzPatrick

Go to the profile of Nature Communications
Sep 24, 2018
0
0

Received Date: 15th August 2018

Purpose:   We aimed to develop an efficient, flexible, scalable and evidence-based approach to sequence-based diagnostic analysis/re-analysis of conditions with very large numbers of different causative genes. We then wished to define the expected rate of plausibly causative variants coming through strict filtering in control in comparison to disease populations to quantify background diagnostic “noise”. 
Methods: We developed G2P (www.ebi.ac.uk/gene2phenotype) as an online system to facilitate the development, validation, curation and distribution of large-scale, evidence-based datasets for use in diagnostic variant filtering.  Each locus-genotype-mechanism-disease-evidence thread (LGMDET) associates an allelic requirement and a mutational consequence at a defined locus with a disease entity and a confidence level and evidence links.  We then developed an extension to Ensembl Variant Effect Predictor (VEP), VEP-G2P, which can filter based on G2P other widely used gene panel curation systems.  We compared the output of disease-associated and control whole exome sequence (WES) using Developmental Disorders G2P (G2PDD; 2044 LGMDETs) and constitutional cancer predisposition G2P (G2PCancer; 128 LGMDETs). 
Results: We have shown a sensitivity/precision of 97.3%/33%  and 81.6%/22.7% for causative de novoand inherited variants respectively using VEP-G2PDDin DDD study probands WES.  Many of the apparently diagnostic genotypes “missed” are likely false-positive reports with lower minor allele frequencies and more severe predicted consequences being diagnostically-discriminative features.  
Conclusion:. Case:control comparisons using VEP-G2PDDestablished an observed:expected ratio of 1:30,000 plausibly causative variants in proband WES to ~1:40,000 reportable but presumed-benign variants in controls. At least half the filtered variants in probands represent background “noise”. Supporting phenotypic evidence is, therefore, necessary in genetically-heterogeneous disorders.  G2P and VEP-G2P provides a practical approach to optimize disease-specific filtering parameters in diagnostic genetic research.

Read in full at bioRxiv.

This is an abstract of a preprint hosted on an independent third party site. It has not been peer reviewed but is currently under consideration at Nature Communications.

Go to the profile of Nature Communications

Nature Communications

Nature Research, Springer Nature