# google/deepvariant

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

Repository: https://github.com/google/deepvariant
Canonical: https://ross.abutalabs.com/products/deepvariant
Language: Python
License: BSD-3-Clause
License Family: permissive
Topics: tensorflow, deep-neural-network, genomics, science, dna, sequencing, genome, bioinformatics, deep-learning, ngs, deepvariant, machine-learning
Last push: 2026-03-19T04:57:17+00:00

## Health v2 (maintenance only)
Score: 70/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 73, release rhythm 49, longevity 100
- inputs: {"age_days": 3206, "days_push": 167, "days_rel": 181, "gap_med": 148, "n_releases_24m": 4}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3791, forks 798 (observed 2026-08-28T04:08:18.706047+00:00)

## What it is
DeepVariant is a deep learning-based genomic variant caller that converts aligned DNA sequencing reads (BAM/CRAM) into pileup image tensors and classifies them with a convolutional neural network to produce VCF/gVCF outputs. It supports germline variant calling in diploid organisms across multiple sequencing platforms including Illumina, PacBio HiFi, Oxford Nanopore, and Complete Genomics.

## Use cases
- call genetic variants from whole genome sequencing data
- identify SNPs and indels from PacBio HiFi long reads
- variant calling on Oxford Nanopore sequencing data
- analyze whole exome sequencing for germline variants
- run trio-based variant calling with DeepTrio
- process RNA-seq data for variant detection
- perform pangenome-aware variant calling

## When to choose
- you need highly accurate germline variant calling from NGS or long-read sequencing data
- you want a deep learning approach that outperforms traditional statistical variant callers
- you work with human diploid genomes across multiple sequencing platforms
- you need reproducible, benchmark-proven variant calling (GIAB winner)

## When to avoid
- you need somatic (cancer) variant calling - use DeepSomatic instead
- your organism is not diploid or has complex copy-number variation
- you lack computational resources for CNN inference on large genomes
- you need a lightweight statistical caller for quick exploratory analysis

## Facets
- artifact type: application
- maturity: stable
- function: machine-learning, deep-learning, data-science
- domain: bioinformatics, machine-learning, deep-learning
- platform: python
- tags: variant-calling, genomics, dna-sequencing, bioinformatics, cnn, tensorflow, vcf, ngs, linux, docker, gpu

## Member repositories
- google/deepvariant (main) score 70

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:18.706047+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:27:04.221065+00:00, confidence not recorded.
  - readme: https://github.com/google/deepvariant (fetched 2026-08-28T04:08:18.706047+00:00, sha 83abb42218bf)
- Data as of 2026-08-30T08:39:29.467469+00:00.