Skip to content

Data Sources Overview

Custom Panel aggregates gene information from multiple trusted sources, each selected for specific strengths in clinical genomics. This page explains each data source, its purpose, and the rationale for inclusion.

Source Categories

Data sources are organized into three main categories:

  1. Clinical Guidelines - Evidence-based recommendations from professional organizations
  2. Community Curated - Expert-reviewed databases with broad community input
  3. Automated/Commercial - Large-scale databases and commercial test offerings

Clinical Guidelines Sources

ACMG Incidental Findings (SF v3.2)

Purpose: Genes recommended for reporting as incidental findings in clinical sequencing

Rationale: - Evidence-based recommendations from the American College of Medical Genetics - Represents genes with actionable clinical interventions - Updated regularly based on clinical evidence - High clinical utility and established management guidelines

Data Type: - Fixed list of ~95 genes - Categories: Hereditary cancer, cardiovascular, metabolic, etc.

Veto Power: Yes - These genes are always included regardless of other scoring

ClinGen Gene-Disease Validity

Purpose: Curated gene-disease relationships with clinical validity classifications

Rationale: - Rigorous curation process by clinical experts - Standardized classification system (Definitive/Strong/Moderate/Limited) - Focus on clinical validity for genetic testing - Transparent evidence evaluation

Data Type: - Gene-disease pairs with validity classifications - Filtered for cancer/neoplasm phenotypes - Updated via live API

Manual Curation Lists

Purpose: Expert-curated gene lists from clinical laboratories or research groups

Rationale: - Incorporates local expertise and clinical experience - Allows inclusion of emerging evidence - Flexible for specific clinical contexts - Can include genes not yet in public databases

Data Type: - Custom Excel/CSV/text files - User-defined evidence scores

Veto Power: Yes - Manually curated genes override scoring thresholds

Community Curated Sources

PanelApp (Genomics England)

Purpose: Community-curated gene panels with evidence-based ratings

Rationale: - Transparent review process by multiple experts - Traffic light system (Green/Amber/Red) indicates confidence - Widely used in clinical practice - Regular updates based on new evidence

Data Type: - Multiple cancer-related panels - Evidence levels: Green (3), Amber (2), Red (1) - Includes MOI and phenotype information

Configuration: 

data_sources:
  PanelApp:
    panels:
      - id: 1    # Childhood solid tumours
      - id: 2    # Adult solid tumours  
      - id: 3    # Haematological malignancies

In-house Panels

Purpose: Gene panels currently used in clinical practice by the laboratory

Rationale: - Reflects current clinical practice - Validated for local patient populations - Incorporates institutional expertise - Ensures continuity of care

Data Type: - Excel/CSV files with gene lists - Can include additional metadata

TheGenCC Database

Purpose: Gene-disease associations from multiple curation efforts

Rationale: - Aggregates data from multiple expert curation groups - Standardized evidence classifications - Broader coverage than single databases - Quality control through member organizations

Data Type: - Gene-disease associations with classifications - Filtered for cancer-related diseases

Scientific Databases

COSMIC Cancer Gene Census

Purpose: Catalog of genes with mutations causally implicated in cancer

Rationale: - Comprehensive cancer gene resource - Tier system indicates evidence strength - Includes both somatic and germline mutations - Widely cited in cancer research

Data Type: - Genes with cancer driver mutations - Tier 1: Well-documented cancer genes - Tier 2: Emerging evidence

Setup Required: COSMIC credentials

HPO/OMIM Neoplasm Genes

Purpose: Genes associated with neoplasm phenotypes in Human Phenotype Ontology

Rationale: - Comprehensive phenotype-gene relationships - Hierarchical ontology captures related concepts - Links to OMIM for detailed clinical information - Useful for discovering less common associations

Data Type: - Genes linked to HP:0002664 (Neoplasm) and descendants - Requires OMIM genemap2.txt file

Setup Required: OMIM access token

Commercial Sources

Commercial Gene Panels

Purpose: Genes included in commercial hereditary cancer test panels

Rationale: - Reflects market adoption and clinical demand - Indicates genes considered actionable by multiple labs - Helps identify consensus across providers - Useful for comparison with academic sources

Included Providers (14 total): - Myriad Genetics - myRisk - Blueprint Genetics - Multiple cancer panels - Invitae - Comprehensive cancer panel - GeneDx - Hereditary cancer panel - Fulgent Genetics - Comprehensive panel - And 9 others...

Data Collection: Web scraping system

Source Selection Criteria

Sources were selected based on:

  1. Clinical Validity - Evidence for gene-disease relationships
  2. Update Frequency - Regular updates with new evidence
  3. Transparency - Clear methodology and evidence criteria
  4. Accessibility - Available data in usable formats
  5. Complementarity - Different perspectives and evidence types

Source Reliability Hierarchy

The scoring system reflects source reliability:

Highest Trust (1.5x weight)
├── ACMG Incidental Findings

High Trust (1.2x weight)  
├── ClinGen
├── TheGenCC
├── In-house Panels
└── Manual Curation

Standard Trust (1.0x weight)
└── PanelApp

Moderate Trust (0.8-0.9x weight)
├── COSMIC (0.9x)
└── Commercial Panels (0.8x)

Lower Trust (0.7x weight)
└── HPO/OMIM Automated

Veto Functionality

The veto system allows critical sources to override the normal scoring threshold:

How It Works

  1. Normal Scoring: Genes must meet the score threshold (default 1.5)
  2. Veto Override: Genes from veto sources are included regardless of score
  3. Clinical Priority: Ensures critical genes aren't excluded

Sources with Veto Power

  • ACMG Incidental Findings: Clinical guidelines override scoring
  • Manual Curation: Expert judgment overrides algorithm

Configuration

data_sources:
  ACMG_Incidental_Findings:
    veto:
      enabled: true
      reason: "ACMG recommended for reporting"

  Manual_Curation:
    veto:
      enabled: true  
      reason: "Expert clinical review"

Use Cases

  • Emerging Genes: Include genes with limited evidence but clinical importance
  • Local Practice: Ensure continuity with current clinical offerings
  • Guidelines Compliance: Meet professional recommendations

Adding New Sources

To add a new data source:

  1. Evaluate against selection criteria
  2. Implement data fetcher in custom_panel/sources/
  3. Configure in default_config.yml
  4. Document rationale and data type
  5. Test integration with existing sources

Next Steps