| name | unicore |
| description | Unicore — universal structure-based core-gene phylogeny using Foldseek and ProstT5 (Steinegger lab). Use when: (1) Building phylogenies from proteomes via structural core genes, (2) 3Di MSAs for tree inference, (3) Scaling beyond sequence-only marker sets. Upstream: https://github.com/steineggerlab/unicore. Route via phylogenomics; depends on foldseek (± GPU ProstT5).
|
| license | MIT |
| category | analysis-tools |
| tags | ["Unicore","Foldseek","ProstT5","phylogeny","core-genes","3Di"] |
| upstream | https://github.com/steineggerlab/unicore |
| stage | taxonomy |
Unicore
Upstream: steineggerlab/unicore
Citation
Kim, D., Park, S. & Steinegger, M. Unicore enables scalable and accurate
phylogenetic reconstruction with structural core genes. Genome Biol. Evol.
17, evaf109 (2025). https://doi.org/10.1093/gbe/evaf109
See also docs/references.md.
Analytical thinking
Unicore finds structural core genes across proteomes (Foldseek clustering /
ProstT5 3Di) and builds phylogenies — a Steinegger-stack alternative/complement
to sequence-marker workflows (phylophlan + iqtree).
| vs | Prefer Unicore when |
|---|
phylophlan / iqtree | Want structure/3Di core genes |
gtdbtk | Custom structural phylogeny, not only GTDB labels |
lucaphylo | Proteome-scale core-gene trees vs single-protein PLM placement |
GPU: --gpu with Foldseek≥10 / ProstT5 on Turing+ NVIDIA (see upstream).
How to run
conda install -c bioconda unicore
unicore easy-core example/data example/results weights tmp
unicore easy-core --gpu example/data example/results weights tmp
Decision tree
Phylogeny from many proteomes/MAGs?
├─ GTDB taxonomy only → gtdbtk
├─ Sequence markers → phylophlan → iqtree
├─ Structural core genes / 3Di → unicore
└─ Single divergent protein family → lucaphylo / foldmason + iqtree
Related skills
phylogenomics · foldseek · foldmason · phylophlan · iqtree ·
gtdbtk · tool-selection
