Summary
Protein language models can predict zero-shot which proteins belong to the same species (1). This is used to cheat when predicting whether two proteins interact.
1.
Hallee L, Peleg T, Rafailidis N, Gleghorn JP. Protein Language Models are Accidental Taxonomists. openRxiv; 2025. Available from: https://doi.org/10.1101/2025.10.07.681002