Variant frequency and/or pathogenicity annotations missing

[julesjacobsen]

In some configurations where a frequency or pathogenicity filter is run after the initial load and filter step then the annotations for these are not present and the variants will not be filtered correctly.

e.g.

  steps:
# Group 1 loads and filters variants
  - failedVariantFilter: {}
  - variantEffectFilter:
      remove:
      - "CODING_TRANSCRIPT_INTRON_VARIANT"
      - "FIVE_PRIME_UTR_EXON_VARIANT"
      - "THREE_PRIME_UTR_EXON_VARIANT"
      - "FIVE_PRIME_UTR_INTRON_VARIANT"
      - "THREE_PRIME_UTR_INTRON_VARIANT"
      - "NON_CODING_TRANSCRIPT_INTRON_VARIANT"
      - "UPSTREAM_GENE_VARIANT"
      - "DOWNSTREAM_GENE_VARIANT"
      - "INTERGENIC_VARIANT"
      - "REGULATORY_REGION_VARIANT"
# Group 2 prioritises and filters genes 
  - hiPhivePrioritiser:
      runParams: "human, mouse, fish, ppi"
  - priorityScoreFilter:
      priorityType: "HIPHIVE_PRIORITY"
      minPriorityScore: 0.501
# Group 3 filters variants - these will NOT work correctly as the annotations will be missing - BUG!!! 
  - frequencyFilter:
      maxFrequency: 1.0
  - pathogenicityFilter:
      keepNonPathogenic: true
# Group 4 MOI-dependent steps (these will always run last as they are dependent on prior filtering steps)
  - inheritanceFilter: {}
  - omimPrioritiser: {}

This will result in missing annotations for variant frequency and pathogenicity scores.

Moving the frequencyFilter up before the hiPhivePrioritiser (i.e. into group 1)

  steps:
# Group 1 loads and filters variants
  - failedVariantFilter: {}
  - variantEffectFilter:
      remove:
      - "CODING_TRANSCRIPT_INTRON_VARIANT"
      - "FIVE_PRIME_UTR_EXON_VARIANT"
      - "THREE_PRIME_UTR_EXON_VARIANT"
      - "FIVE_PRIME_UTR_INTRON_VARIANT"
      - "THREE_PRIME_UTR_INTRON_VARIANT"
      - "NON_CODING_TRANSCRIPT_INTRON_VARIANT"
      - "UPSTREAM_GENE_VARIANT"
      - "DOWNSTREAM_GENE_VARIANT"
      - "INTERGENIC_VARIANT"
      - "REGULATORY_REGION_VARIANT"
  #  frequencyFilter moved into group 1. Annotations now present 
  - frequencyFilter:
      maxFrequency: 1.0
# Group 2 prioritises and filters genes 
  - hiPhivePrioritiser:
      runParams: "human, mouse, fish, ppi"
  - priorityScoreFilter:
      priorityType: "HIPHIVE_PRIORITY"
      minPriorityScore: 0.501
# Group 3 filters variants - these will NOT work correctly as the annotations will be missing - BUG!!! 
  - pathogenicityFilter:
      keepNonPathogenic: true
# Group 4 MOI-dependent steps (these will always run last as they are dependent on prior filtering steps)
  - inheritanceFilter: {}
  - omimPrioritiser: {}

n.b. the above configuration is NOT recommended and will produce incorrect results in versions <=13.3.0

[julesjacobsen]

Fixed in version 14.0.0