IMPACT-SNV

Integrated Mapping of Phenotype-Associated Candidate Targets for SNV/Indel Analysis

This repository contains the IMPACT-SNV pipeline, which processes and prioritizes single nucleotide variants (SNVs) and indels for rare disease analysis using the FAVOR database and phenotype-specific gene-disease associations.

Overview

IMPACT-SNV is part of the broader IMPACT framework for phenotype-configurable interpretation of genomic variants. This module specifically handles SNV/Indel processing and produces output files compatible with IMPACT-VIS for interactive visualization and analysis.

Pipeline Architecture

The pipeline consists of four modular steps, each available as a DNAnexus applet:

VCF Files → [Step 1: Merge] → [Step 2: VCF2GDS] → [Step 3: FAVOR Annotate] → [Step 4: Prioritize] → *_SNV_IMPACT.gds

Step	Folder	Description	Input	Output
1	step1_vcf_merge/	Merges multiple VCF files into chromosome-separated files	.vcf, .vcf.gz	merged_chr*.vcf.gz
2	step2_vcf2gds/	Converts VCF to GDS format for efficient processing	.vcf.gz	merged_chr*.gds
3	step3_favorannotator-rap/	Annotates variants using FAVOR database	.gds	favor_merged_chr*.gds
4	step4_impact_prioritization/	Scores and prioritizes variants by pathogenicity	.gds, GeneList.txt	*_SNV_IMPACT.gds

Requirements

Software Dependencies

• R ≥ 4.0 with Bioconductor packages:
  • SeqArray, SeqVarTools, gdsfmt
  • stringi, dplyr, tidyr
• Python 3 (for VCF merging scripts)
• BCFtools, bgzip, tabix (for VCF processing)
• Rust and xsv (for FAVOR annotation step)

Platform

These tools are designed for deployment on the DNAnexus Research Analysis Platform. Each folder contains a dxapp.json configuration file for building DNAnexus applets.

Input Requirements

Step 1: VCF Merge

• Input: One or more VCF or VCF.GZ files
• Output: Chromosome-separated VCF.GZ files (merged_chr1.vcf.gz, merged_chr2.vcf.gz, etc.)

Step 2: VCF to GDS

• Input: VCF.GZ file from Step 1
• Output: GDS file (merged_chr*.gds)

Step 3: FAVOR Annotation

• Input: GDS file from Step 2
• Output: Annotated GDS with FAVOR functional annotations

Step 4: IMPACT Prioritization

• Input:
  • Annotated GDS files from Step 3 (merged_chr*.gds)
  • Gene-disease association file (GeneList.txt) - tab-separated with columns:
    • symbol: Gene symbol (e.g., GJB2, OTOF)
    • globalScore: Open Targets association score (0-1)
• Output: Per-sample GDS files ({sample_id}_SNV_IMPACT.gds)

Output File Format

The final *_SNV_IMPACT.gds files contain:

Core SeqArray Nodes

Node	Type	Description
variant.id	integer	Unique variant identifier
chromosome	integer/character	Chromosome
position	integer	1-based genomic position
sample.id	character	Sample identifiers
genotype	integer	Genotype array

IMPACT Score Annotations

Node	Type	Description
annotation/info/impact_score	numeric	Pathogenicity score (0-100)
annotation/info/impact_score_calc	character	Tier and calculation formula
annotation/info/tier	integer	Priority tier (1-4)

ClinVar Significance Flags

Boolean indicators under annotation/info/clnsig_flags/:

• pathogenic, likely_pathogenic, uncertain_significance
• likely_benign, benign
• pathogenic_low_penetrance, likely_pathogenic_low_penetrance
• established_risk_allele, likely_risk_allele, uncertain_risk_allele
• affects, association, drug_response, confers_sensitivity
• protective, other, conflicting_interpretations_of_pathogenicity, not_provided

FAVOR Functional Annotations

Node	Description
annotation/info/FunctionalAnnotation/VarInfo	Functional consequence
annotation/info/FunctionalAnnotation/genecode_comprehensive_info	Gene information
annotation/info/FunctionalAnnotation/clnsig	ClinVar clinical significance
annotation/info/FunctionalAnnotation/clndn	ClinVar disease name
annotation/info/FunctionalAnnotation/bravo_af	Bravo allele frequency
annotation/info/FunctionalAnnotation/apc_protein_function_v3	Protein function score

Usage

DNAnexus Platform

Step 1: VCF Merge

dx run step1_vcf_merge \
  -ivcfs=sample1.vcf.gz \
  -ivcfs=sample2.vcf.gz \
  -o merged_output

Step 2: VCF to GDS Conversion

dx run vcf2gds \
  -ivcf_file=merged_chr1.vcf.gz \
  -igds_filename=merged_chr1.gds

Step 3: Functional Annotation

dx run favorannotator \
  -igds=merged_chr1.gds \
  -ichromosome=1 \
  -iuse_compression=TRUE

Step 4: Variant Prioritization

dx run step4_impact_prioritization \
  -igenelist=GeneList.txt \
  -igds_files=favor_merged_chr1.gds \
  -igds_files=favor_merged_chr2.gds

Local Execution

For Step 4 local execution:

cd step4_impact_prioritization/resources/home/dnanexus/
Rscript IMPACT-prioritization.r --genelist GeneList.txt --outprefix anno_merged_

Integration with IMPACT-VIS

The output *_SNV_IMPACT.gds files are designed for direct use with IMPACT-VIS, an interactive R Shiny application for variant visualization.

Preparing Data for IMPACT-VIS

1. Place output files in the IMPACT-VIS app/data/ directory
2. Follow the naming convention: {sample_id}_SNV_IMPACT.gds
3. See the IMPACT-VIS Data Preparation Guide for details

IMPACT-VIS Capabilities

• Interactive visualization of prioritized variants
• Filtering by tier, ClinVar significance, allele frequency
• Gene-based and genomic region filtering
• Publication-ready plots with Plotly
• Persistent annotation states for sample curation

Tiering System

Variants are assigned to tiers based on evidence strength:

Tier	Criteria	Score Formula
Tier 1	Pathogenic/Likely Pathogenic in ClinVar + gene match	80 + 20 × globalScore
Tier 2	Frameshift, stopgain mutations	60 + 40 × globalScore
Tier 3	Nonsynonymous, nonframeshift, stoploss	20 + 80 × globalScore
Tier 4	APC protein function evidence	100 × (0.5 × APC + 0.5 × globalScore)

Gene-Disease Association File

IMPACT-SNV requires a phenotype-specific gene-disease association file. This file is supplied by the user and is not generated automatically by IMPACT-SNV during runtime.

For the analyses described in the IMPACT manuscript, gene lists were manually generated from Open Targets by querying the phenotype of interest, exporting associated genes, and formatting the results as a tab-separated text file.

The file must contain the following two columns:

• symbol: approved gene symbol
• globalScore: Open Targets global gene-disease association score for the queried phenotype

Example:

symbol    globalScore
GJB2    0.858985237
OTOF    0.850795742
MYO6    0.845566359

To generate this file:

1. Query Open Targets for the phenotype or disease of interest.
2. Export the associated genes and global association scores.
3. Retain or rename the gene-symbol column as symbol.
4. Retain or rename the Open Targets global association score column as globalScore.
5. Save the file as a tab-separated text file with a header row.

Users may also supply manually curated phenotype-specific gene panels or gene lists from other resources. In that case, the file must still contain both symbol and globalScore. If using manually curated gene panels or non-Open Targets resources, users must provide a numeric globalScore column. We recommend documenting how these values were assigned or normalized, as they contribute to IMPACT-SNV prioritization.

References

Tools and Databases

• SeqArray - Efficient storage of sequence data
• FAVOR - Functional Annotation of Variants Online Resource
• Open Targets - Gene-disease association platform
• DNAnexus - Cloud-based genomics platform

Related IMPACT Modules

• IMPACT-VIS - Interactive visualization
• IMPACT-SV - Structural variant processing
• IMPACT-CNV - Copy number variant processing

Citation

If you use this pipeline, please cite:

@article{impact-TBA,
  title={Integrated Mapping of Phenotype-Associated Candidate Targets for 
         interpretation and prioritization of genomic variants},
  authors={Boehler, N. and Cheng, H. Y. M.},
  journal={TBA},
  year={TBA}
}

FAVOR Citation

Zhou H., et al. (2023). FAVOR: functional annotation of variants online resource and annotator for variation across the human genome. Nucleic Acids Research, 51(D1), D1300-D1311. DOI: 10.1093/nar/gkac966

License

This project is licensed under the terms specified in the LICENSE file.

Support

• Issues: Report bugs or request features on the GitHub Issues page
• Documentation: See individual step READMEs for detailed usage
• IMPACT-VIS Help: Visit the IMPACT-VIS documentation