Virtual panel parser for Cellbase annotated variants. The aim is to replicate previous work, which used a list of gene symbols and a rule-set to return a list of possible pathogenic variants from Alamut annotated VCF data variant_parser.
This implementation will include annotation via Cellbase from the original VCF. To keep connections to various RESTful services down the gene list is used to pre-filter the VCF based on the genomic co-ordinates returned from ENSEMBL Biomart.
Population frequency data should then be retrieved from Gnomad via Tabix, and the rule-set applied with the final output being a list of candidate variants with additional annotation and reason for shortlisting.
Uses HTSJDK to parse VCF files and submit the genomic coordinates to the Cellbase RESTful service for annotation.
Outline plan:
- Ability to import VCF
- Query Cellbase for annotation
- Deconstruct returned JSON using Odin
- Filter by ruleset and provided gene symbol list - check list against HGNC
- Score and order shortlist by impact from Cellbase, population frequency from Gnomad
Additional future developments:
- Reactome integration - retrieve direct interacts with gene symbol products and use these to increase the size of the gene list
- Support for CNV calls - must be included as standard VCF (no custom data formats)
- Look at making ruleset configurable by introducing a DSL
- PanelApp / Monarch API integration - allow the user to provide a phenotype(s) and return the related gene symbols
- Support for multi-sample VCF and segregation analysis - will require a PED file for family structure
Based on Katherine Smith's presentation of Genomics England Rare Disease Filtering rules (ACGS Exome Workshop 2017) Rules adapted were required:
- Criterion 1: Variant must contain PASS filter (this maybe flagged to allow criterion to be relaxed)
- Criterion 2: Allele frequency 0.01 Biallelic / 0.001 Monoallelic (can be set by args)
- Criterion 3:
- Following Biotypes allowed: protein_coding; nonsense_mediated_decay; non_stop_decay; IG_C_gene; IG_D_gene; IG_J_gene; IG_V_gene; TR_C_gene; TR_D_gene; TR_J_gene; TR_V_gene
- High consequence: transcript_ablation; splice_acceptor_variant; splice_donor_variant; stop_gained; frameshift_variant; stop_lost; initiator_codon_variant
- Moderate consequence: transcript_amplification; inframe_insertion; inframe_deletion; missense_variant; splice_region_variant; incomplete_terminal_codon_variant
- Criterion 4: Segregation (maybe future work since it will require a pedigree)
- Criterion 5: Gene panels (checked against HGNC api)
Copyright © 2017 Garan Jones
Distributed under the Eclipse Public License either version 1.0 or (at your option) any later version.