Can it handle methylation calling?

Yes, it supports 5mCG_5hmCG modifications and automatically adjusts memory and batch size for these intensive tasks.

What basecalling models are available?

It supports fast, hac (high accuracy), and sup (super high accuracy) model tiers with automated resource calculation for each.

Which HPC clusters does this skill support?

It is specifically configured for University of Michigan's ARMIS2 and Great Lakes clusters, including specific partition and account defaults.

How does it track experimental data?

It integrates with the ont-experiments framework to capture model paths, GPU allocations, BAM statistics, and job metadata.

Dorado ONT Basecalling Bench

Name: Dorado ONT Basecalling Bench
Author: Single-Molecule-Sequencing

bySingle-Molecule-Sequencing

0•

数据科学与机器学习

Automates Oxford Nanopore basecalling workflows on HPC clusters with integrated provenance tracking and GPU optimization.

Dorado-Bench v2 is a specialized tool for managing Oxford Nanopore (ONT) basecalling tasks specifically optimized for University of Michigan's ARMIS2 and Great Lakes HPC clusters. It streamlines the generation of SLURM jobs, manages GPU resource allocation for different model tiers (fast, hac, sup), and integrates with the ont-experiments framework to ensure full experimental provenance. Whether you are processing POD5 data, performing methylation calling, or benchmarking model performance, this skill automates the complex resource calculations and metadata capturing required for high-throughput genomic data processing.

主要功能

01Automated SLURM job generation for UM HPC clusters (ARMIS2 and Great Lakes)

02Integrated provenance tracking through ont-experiments for experimental reproducibility

030 GitHub stars

04Dynamic resource allocation for fast, HAC, and SUP model tiers

05Native support for methylation calling with automatic resource adjustments

06Comprehensive GPU optimization and metadata capturing for BAM statistics

使用场景

01Ensuring research compliance through automated provenance and metadata tracking

02Benchmarking Dorado model performance across different GPU hardware configurations

03Running high-accuracy basecalling on large-scale Nanopore datasets using UM clusters

主要功能

01Automated SLURM job generation for UM HPC clusters (ARMIS2 and Great Lakes)

02Integrated provenance tracking through ont-experiments for experimental reproducibility

030 GitHub stars

04Dynamic resource allocation for fast, HAC, and SUP model tiers

05Native support for methylation calling with automatic resource adjustments

06Comprehensive GPU optimization and metadata capturing for BAM statistics

使用场景

01Ensuring research compliance through automated provenance and metadata tracking

02Benchmarking Dorado model performance across different GPU hardware configurations

03Running high-accuracy basecalling on large-scale Nanopore datasets using UM clusters