Introduction
Myloasm is a de novo metagenome assembler for long-read sequencing data. It takes sequencing reads and outputs polished contigs in a single command.
Myloasm works with modern long reads such as:
- Nanopore R10 simplex reads with > ~97% accuracy (basecalled in sup or hac mode)
- PacBio HiFi reads
Installation and Usage: See the Install and Usage sections on the navigation side bar. Source code is available at the GitHub repo.
Announcement
Myloasm is now published in Nature Biotechnology. (March 27, 2026)
Why myloasm?
Results: Here are some preliminary results for myloasm; see the preprint for more details.
Philosophy: Myloasm was designed to take advantage of modern long reads. Even the noisiest modern long reads (e.g., nanopore simplex R10) have become quite accurate. Myloasm uses a new algorithmic framework that enables high-resolution assembly from this data.
Strengths:
- leverages modern long-reads (e.g. nanopore R10.4) to get longer, better contigs than other assemblers
- assembles similar (intraspecies) strains better than other nanopore assemblers
- other nanopore assemblers collapse similar genomes at > 95% ANI—myloasm can resolve up to 99% ANI quite confidently.
- works well in diverse metagenomes, from human gut to soil
- uses about the same RAM as metaFlye with faster runtime
Limitations: myloasm may
- occasionally produce chimeric misassembled contigs (like all metagenome assemblers)
- we provide tools for quality control and visualization; also see the quality control guide.
- use more RAM than sparse minimizer approaches like metaMDBG.
Issues, questions, and discussions
- Found a bug or have an issue? Go to https://github.com/bluenote-1577/myloasm/issues
- Have a general question or discussion topic? Go to https://github.com/bluenote-1577/myloasm/discussions