BITTENSOR SUBNET

Turning DNA analysis into an open competition

GenotypeNet transforms variant calling—the foundational layer of genomics—into a public proving ground where anyone can compete.

We create realistic genome challenges. Validators score submissions using established benchmarking standards. Everything is public — it's clear what works, what doesn't, and how methods truly compare and improve over time.

How it works

Synthetic Genomes

Cryptographically unique

We generate genomes with carefully selected, biologically plausible mutations, unique to each round. ~3.2 billion positions—impossible to game.

Gold-Standard Truth Genotypes

Validators know ground truth

Each task is randomized and truth-checked. Only real genomic algorithms can win.

Run Your Pipeline

Any tool, any approach

Miners download the task, run any tools they choose, and submit their algorithms.

Automated Scoring

hap.py standard

Validators compare calls to truth using industry-standard benchmarking and publish scores.

Open Competition

Results speak

Open-source scoring creates a fair proving ground. Best methods rise to the top.

Why it matters

The Foundation

Variant calling is the foundational step. Everything downstream depends on this.

Real Impact

Better calls enable clearer diagnostics and faster research.

Honest Progress

Public scoring keeps evaluation honest. No cherry-picked benchmarks.

Who it's for

Bioinformatics teamsvalidating pipelines
ML researchersbuilding models
Statisticianstesting hypotheses
Engineersexploring algorithms
Learnersunderstanding DNA

Ready to compete?

Join the subnet, run your algorithm, and watch your performance go live.