MMseqs2-GPU sets new standards in single query search speed, allows near instant search of big databases, scales to multiple GPUs and is fast beyond VRAM. It enables ColabFold MSA generation in seconds and sub-second Foldseek search against AFDB50. 1/n
📄 www.nature.com/articles/s41...
💿 mmseqs.com

Spacedust is a collaborative effort with the amazing @milot.bsky.social and Johannes Soeding.
6/6

Spacedust is available at
🚀 github.com/soedinglab/s...
Install Conda Package via
🐍 conda install -c bioconda spacedust
5/6

Spacedust recovers previously annotated gene clusters e.g. operons, antiviral defense systems, and BGCs. It also identifies several more instances of CRISPR subtype III-E in the GTDB. 4/6

We searched all-vs-all 1308 bacterial genomes from different genera. 1) Spacedust assigns 58% of all 4.2M genes & 35% of the unannotated genes into conserved gene clusters. 2) It offers better functional association prediction based on the congruence of KEGG module IDs.
3/6

Spacedust finds all gene clusters significantly conserved between any two genomes in a set of input genomes, by iteratively merging smaller clusters to maximize the statistical significance of the degree of clustering and order/strand conservation, therefore being reference-free.
2/6

Happy to see that Spacedust is now published on Nature Methods!
It combines sensitive Foldseek structure search and conserved neighborhood detection to discover functionally-associated gene clusters in prokaryotic & viral genomes.
1/6🧵

Spacedust finds all gene clusters significantly conserved between any two genomes in a set of input genomes, by iteratively merging smaller clusters to maximize the statistical significance of the degree of clustering and order/strand conservation, therefore being reference-free.
2/6