deepvariant

DeepVariant is an analysis pipeline that uses a deep neural network to call genetic variants from next-generation DNA sequencing data.

Whole genome sequencing analysis pipeline for consumer hardware. 100% local, Docker-powered, free and open source.

DeepVariant-on-Spark is a germline short variant calling pipeline that runs Google DeepVariant on Apache Spark at scale.

Scalable and High Performance Variant Calling on Cluster Environments

Joint variant calling with GATK4 HaplotypeCaller, Google DeepVariant 1.0.0 and Strelka2, coordinated via Snakemake.

Notebooks and examples for DeepVariant on Spark project.

A Snakemake workflow for variant calling with DeepVariant, and optionally joint variant calling using GLnexus.

Workflows for whole-genome/exome sequencing data analysis

Synthetic somatic mutation read-generation, variant calling pipelines (DeepVariant & GATK), and benchmarking framework for bioinformatics research.

Benchmarking parabricks GPU-accelerated tools for GATK and beyond!

Local, privacy-first analyser for your own whole-genome sequencing data: FASTQ to GRCh38 calls with DeepVariant on your own machine, graded against ClinVar, gnomAD and ClinGen, with pharmacogenomics (PharmCAT, CPIC), polygenic scores, haplogroups and repeat expansions. Nothing leaves your machine unless you ask; a monthly refresh shows what's new.

Event-driven AWS genomics pipeline: FASTQ → VCF with bwa-mem2, DeepVariant and pangenome-aware calling. Terraform, Step Functions, Batch; benchmarked on GIAB.

Master's Thesis Project

Native Rust reimplementation of DeepVariant 1.10.0 with CPU, Metal, CUDA and ROCm inference (byte-identical make_examples target). Work in progress. Not the official DeepVariant.

Release validation for a germline WGS workflow (WDL): GATK and DeepVariant benchmarked against GIAB HG002, with a regression gate, case studies and runbooks

Nextflow pipeline for variant discovery from PacBio HiFi reads, with phasing, optional SV/methylation and VEP annotation.

Self-built GPU-accelerated WGS pipeline (NVIDIA Parabricks fq2bam + DeepVariant). Runs on a laptop GPU via Docker or a multi-GPU cluster via rootless Podman. hg38 + T2T, checkpoint-resume.

Pipeline Nextflow DSL2 de variant calling germinal sur chr20, benchmarke contre le truth set GIAB HG002 avec hap.py. GATK, bcftools et DeepVariant compares sur donnees identiques.

Improved whole-genome variant calling in the green sea turtle (Chelonia mydas) with a DeepVariant model retrained on a simulation-derived truth set: code for the BMC Genomics manuscript

Analyze whole genome sequencing data on consumer hardware with no cloud accounts, subscriptions, or bioinformatics degree needed