Complex Portal expands Human Complexome with ML-predicted human protein complexes

Complex Portal 248 release is now publicly available to explore

The Complex Portal – an encyclopaedic resource of macromolecular complexes from key model organisms – has expanded its coverage of the Human Complexome. This includes:

• 2,228 curated human complexes, plus complexes from many other species
• 14,964 new machine-learning predicted complexes from hu.MAP3.0.
  • High-confidence assemblies predicted on large-scale experimental data.
  • Training based on 2150 high quality manually curated human Complex portal complexes.
  • Explore hu.MAP3.0: Atlas of human protein complexes by integration of > 25,000 proteomic experiments for further details.

• A new Complex Navigator tool available to facilitate comparison of complexes. This can be done by species, participant type or by orthologous and/or paralogous complexes

• Filter options to choose between curated and predicted complexes.

To find out more, read the new paper entitled Complex Portal 2025: Predicted human complexes and enhanced visualisation tools for the comparison of orthologous and paralogous complexes published in Nucleic Acids Research (NAR).

---

Source article(s)

Complex portal 2025: predicted human complexes and enhanced visualisation tools for the comparison of orthologous and paralogous complexes

Balu S. et al.

Nucleic Acids Research 18 November 2024

10.1093/nar/gkae1085

Tags: embl-ebi, FAIR data, machine learning, protein complex