Open access
Sep 2026
ContrasTED: contrastive domain embeddings for scalable remote homology classification
ContrasTED uses CATH-supervised center-contrastive learning to project structure-aware embeddings into a domain-level metric space for nearest-centroid superfamily assignment, which separates superfamilies while retaining structural information below 20% sequence identity.
David Miller, Nicola Bordin, Janusan Jeyananthan et al.
· bioRxiv · 0 citations