Command-line tool that accepts a TSV file of variants by samples and emits an XLSX file of genes by samples, rolling up one or more variants into a single gene row.
The official repository is at:
https://github.com/umich-brcf-bioinf/GeneRollup
For help and a description of command-line options, run:
$ generollup -h
There are some constraints on the input file:
- It is a tab-separated file.
- Each row represents a single variant locus associated with a gene. (Recommend that you filter the input to exclude intergenic regions.)
- Each row represents a single annotation, typically the "worst" impact. Since each locus could be associated with more than one alt allele and/or multiple annotations, you may need to process the variant calling output to limit to the annotation of choice.
- The column headers must include
- gene symbol
- sample genotype columns
- some combination of variant annotation
- SnpEff: Impact (High,Moderate,Low,Modifier)
- dbNSFP: count of annotation sources that designated this variant as damaging
- VarSeq (GoldenHelix)
Here is toy example input file to illustrate the syntax. This example assumes you have annotated with dbNSFP and SnpEff:
| #CHROM | POS | REF | ALT | GENE_SYMBOL | dbNSFP_damaging | SNPEFF_IMPACT | GT.P1.TUMOR | GT.P2.TUMOR |
|---|---|---|---|---|---|---|---|---|
| chr1 | 14948 | G | A | BRCA1 | 1 | HIGH | 0/0 | 0/1 |
| chr1 | 137622 | G | A | TANK | 0 | . | . | 1/1 |
| chr1 | 1147545 | A | G,C | CREBBP | 7 | LOW | 0/2 | 0/0 |
| chr1 | 1147548 | A | T | CREBBP | 7 | MODERATE | . | 0/1 |
$ generollup \
sample_genotype_column_regex='GT\.(.*).TUMOR' \
gene_column_name='GENE_SYMBOL' \
effect_column_name='SNPEFF_IMPACT' \
effect_column_values='HIGH,MODERATE,LOW,MODIFIER' \
dbnsfp_column_name='dbNSFP_damaging' \
input.tsv output.xlsx
See documentation on github for more details on how the input variants are aggregated.
====
Email bfx-jacquard@umich.edu for support and questions.
UM BRCF Bioinformatics Core