⚙️ Exomiser Settings — Analysis Template

What is this file?

This YAML file defines how Exomiser filters, scores, and prioritizes variants.
Think of it as the analysis pipeline configuration — every step from allele-frequency filtering to phenotype prioritization is controlled here.

🧠 Overview

The template below represents a typical PASS_ONLY analysis configuration.

Key concepts:

analysisMode — which variants from the VCF are evaluated
inheritanceModes — weights applied to genetic models
filters — remove unlikely variants
prioritisers — rank genes based on phenotype evidence

🧬 Analysis Mode

PASS_ONLY

Only variants marked as PASS in the VCF are analysed. This avoids low-quality variant calls.

🧬 Inheritance Models

How weighting works

Larger numbers increase the importance of that inheritance model during scoring.

Examples:

AUTOSOMAL_RECESSIVE_COMP_HET = strong weight (2.0)
AUTOSOMAL_DOMINANT = lower weight (0.1)

🌍 Frequency Sources

These datasets help remove common population variants:

UK10K
gnomAD exomes (GNOMAD_E_*)
gnomAD genomes (GNOMAD_G_*)

Note

As of gnomAD v4, several legacy datasets (1000G, ESP, ExAC) are incorporated into gnomAD.

🔬 Pathogenicity Sources

The configuration uses:

REVEL
MVP

CADD / REMM

If you enable CADD or REMM, you must install the tabix files and update their paths in application.properties.

🧩 Analysis Steps Explained

The steps section defines the Exomiser pipeline.

Typical flow:

Variant filtering
Frequency filtering
Inheritance filtering
Phenotype prioritisation

Important

inheritanceFilter and omimPrioritiser should run after all filters.

📄 Full YAML Configuration

Show Analysis Template (click to expand)

## Exomiser Analysis Template.
---
analysisMode: PASS_ONLY
inheritanceModes: {
  AUTOSOMAL_DOMINANT: 0.1,
  AUTOSOMAL_RECESSIVE_HOM_ALT:  0.1,
  AUTOSOMAL_RECESSIVE_COMP_HET: 2.0,
  X_DOMINANT: 0.1,
  X_RECESSIVE_HOM_ALT: 0.1,
  X_RECESSIVE_COMP_HET: 2.0,
  MITOCHONDRIAL: 0.2
}

frequencySources: [
  UK10K,
  GNOMAD_E_AFR,
  GNOMAD_E_AMR,
  GNOMAD_E_EAS,
  GNOMAD_E_NFE,
  GNOMAD_E_SAS,
  GNOMAD_G_AFR,
  GNOMAD_G_AMR,
  GNOMAD_G_EAS,
  GNOMAD_G_NFE,
  GNOMAD_G_SAS
]

pathogenicitySources: [ REVEL, MVP, ALPHA_MISSENSE, SPLICE_AI ]

steps: [
  failedVariantFilter: { },

  variantEffectFilter: {
    remove: [
      FIVE_PRIME_UTR_EXON_VARIANT,
      FIVE_PRIME_UTR_INTRON_VARIANT,
      THREE_PRIME_UTR_EXON_VARIANT,
      THREE_PRIME_UTR_INTRON_VARIANT,
      NON_CODING_TRANSCRIPT_EXON_VARIANT,
      UPSTREAM_GENE_VARIANT,
      INTERGENIC_VARIANT,
      REGULATORY_REGION_VARIANT,
      CODING_TRANSCRIPT_INTRON_VARIANT,
      NON_CODING_TRANSCRIPT_INTRON_VARIANT,
      DOWNSTREAM_GENE_VARIANT
    ]
  },

  frequencyFilter: {maxFrequency: 2.0},
  pathogenicityFilter: {keepNonPathogenic: true},
  inheritanceFilter: {},
  omimPrioritiser: {},
  hiPhivePrioritiser: {}
]

Next Module: Running Exomiser with phenopacket