> ## Documentation Index > Fetch the complete documentation index at: https://docs.theminos.ai/llms.txt > Use this file to discover all available pages before exploring further. # Overview > Minos — Decentralized Genomic Variant Calling on Bittensor Subnet 107 # The Foundational Layer of Genomics Minos is a **decentralized genomic variant-calling network** running on [Bittensor Subnet 107](https://github.com/minos-protocol/minos_subnet). Miners compete to produce the most accurate variant calls from sequencing data. Validators score results against truth sets — **accuracy is the only thing that matters.** **SN107** rewards miners based on the quality of their genomic variant calls. *** ## How Minos Works Validators send miners a genomic region and aligned sequencing reads (BAM) to analyze. Miners run one of the four supported variant callers with their own tuned hyperparameters to produce variant calls (VCF) and share their hyperparameter config with validators for scoring. Validators generate the VCF using the miner's config and benchmark against a truth set using precision, recall, and F1-score. *** ## Supported Variant Callers Minos supports **4 industry-standard tools**. Each guide covers the tool's algorithm, pipeline, and every tunable hyperparameter. The Broad Institute's flagship caller. Performs **local de novo assembly** of haplotypes — Assembly-based approach for haplotype-aware calling. Deep learning variant caller. Converts pileups into **CNN image classification** — platform-aware models for WGS, WES, and PacBio. Bayesian haplotype-based caller with the **most tunable priors** — toggle HWE, allele balance, and contamination models independently. The fastest caller. Classic **pileup-based approach** with a two-stage pipeline giving independent control over data prep and genotyping. *** ## Quick Links How the subnet works end-to-end. Get mining in under 10 minutes. Source code and issue tracker.