nf-cavalier

Nextflow Pipeline for singleton and family based candidate variant prioritisation based on gene lists using the Cavalier R package. This pipeline is a work in progress.

Installation

• Clone this repositoty

Usage

• Create and navigate to run working directory

• Create configuration file in run directory named nextflow.config:

    params {
      // output directory
      outdir = 'output'
      
      // inputs
      snp_vcf = 'my_cohort.SNPs.vcf.gz'
      sv_vcf = 'my_cohort.SVs.vcf.gz'
      ped = 'families.ped'
      bams = 'bams.tsv'
      lists = 'my_list_file.tsv,HP:0001250'
      
      // filtering
      maf_dom = 0.0001
      maf_rec = 0.01
      maf_comp_het = 0.01
      maf_de_novo = 0.0001
      max_cohort_af = 1.0
      max_cohort_ac = 'Inf'
      min_impact = 'LOW'
      exclude_benign_missense = false
      include_sv_csv = true
  
      // reference config
      ref_hg38 = true
      ref_fasta = '/PATH/TO/GRCh38.fasta'
      pop_sv = '/PATH/TO/gnomad-sv.vcf.gz'
      ref_gene = '/PATH/TO/RefSeqGene.hg38.UCSC.txt'
      vep_cache = '/PATH/T0/vep-cache'
      vep_cache_ver = '104'
    }

• Note: Bahlo Lab members should use this config as a starting point.

• Params

  • outdir - Output directory
  • snp_vcf - (Optional) Input VCF file with SNP variant calls for all samples.
  • sv_vcf - (Optional) Input VCF file with SV variant calls for all samples.
  • ped - A Ped format file describing familial relationships, with 1/2 coding for unaffected/affected phenotypes (missing phenotype not supported).
  • bams - TSV file with first column containing individual ID, second column containing path to indexed BAM file (no header row/ column names).
  • lists - Comma separated list of gene lists to use for filtering. This may be a local TSV file, e.g. 'my_list_file.tsv' or a web based gene list, e.g. PAA:289.

    • Local TSV file

      • Path to a TSV file with mandatory named column. The file should have at least one of the following column names: ensembl_gene_id, hgnc_id, entrez_id or symbol. Optional column names are list_id, list_name, list_version and inheritance. Note that all gene IDs are ultimately converted to Ensembl Gene IDs using HGNC to match VEP annotation.

        e.g.

        list_id	list_name	list_version	symbol	inheritance
        PAA:202	Genetic Epilepsy	1.26	AARS1	AR
        PAA:202	Genetic Epilepsy	1.26	ABAT	AR
        PAA:202	Genetic Epilepsy	1.26	ABCA2	AR
        

        or:

        list_id	list_name	list_version	ensembl_gene_id	inheritance
        PAA:202	Genetic Epilepsy	1.26	ENSG00000090861	AR
        PAA:202	Genetic Epilepsy	1.26	ENSG00000183044	AR
        PAA:202	Genetic Epilepsy	1.26	ENSG00000107331	AR
        

    • Web List - Cavalier will automatically retrieve the latest version of these web lists

      • PanelApp: PanelApp Australia or PanelApp Genomics England lists may be specified with "PAA:" or "PAE:" prefix respectively. e.g. PAA:289
      • HPO: Human phenotype ontology terms may be specified with the "HP:" prefex, e.g. HP:0001250
      • Genes4Epilepsy: Genes4Epilepsy lists may be specified with the "G4E:" prexif, e.g. "G4E:ALL" for All Epilepsy genes, or "G4E:Focal" for Focal epilepsy genes only
      • HGNC: Gene subsets by locus group can be extracted from HGNC, for example "HGNC:protein-coding" will give a list of all protein coding genes
  • exclude_benign_missense - exclude missense variants that are predicted/annotated as benign by all of Sift, Polyphen and ClinVar. Missing annotations are ignored.
  • include_sv_csv - Include "coding_sequence_variant" SVs regardless of VEP Impact.

• First run:
  nextflow run /PATH/TO/nf-cavalier

• Resume run:
  nextflow run /PATH/TO/nf-cavalier -resume

• Recommended to run workflow in a screen session