winnowr
winnowr post-processes single-sample GATK HaplotypeCaller VCFs with models you train on truth-labelled samples.
What sets winnowr apart
- Works on a single sample: no cohort or joint calling.
- Fixes calls, not just filters them: it can drop false calls, flip wrong zygosity, re-genotype indels in repeats, and rescue true calls that were filtered.
- Uses evidence the VCF lacks: repeat-tract lengths and reads below HaplotypeCaller’s MAPQ cut-off.
- Tuned for F1: cross-validation against your truth data picks each model’s thresholds.
- Drops in after HaplotypeCaller: no re-calling.
No trained models are distributed: you train them on samples with a truth set, such as GIAB, labelled by hap.py. The training quick start lists the steps, and Classes and labels explains what each action means.
The VCF must be genotyped from a HaplotypeCaller gVCF (e.g. one made with --dragen-mode). score
and extract also take that gVCF (--gvcf), and read it only when the VCF lacks the raw
annotations.
By default score also applies two indel rules that are fixed, not chosen by cross-validation. An
indel call that the thresholds leave alone is dropped when its P(keep) is below 0.5. Given GIAB’s
not-difficult regions (--regions notinalldifficultregions=...), an indel call outside them is
dropped at a threshold of 0.6, in place of the model’s drop threshold. See
Scoring.
Demonstration code only. winnowr is a research prototype, shared to illustrate an approach to machine-learned filtering and genotyping of germline variant calls. It is not supported, has not been validated, and is not intended for clinical, diagnostic or production use. Use it at your own risk.
The commands
winnowr has eight commands: strlen, pileup, extract, subset, train, evaluate,
stratify and score. The tool reference describes each, with its full
--help.