Local structural preference maps encode transferable protein-interface energetics
- 1Helmholtz Munich
Abstract
Relating the drivers of binding affinity to structural features remains difficult because affinity emerges from many weak, context-dependent interactions, while experimental affinity and mutation data are limited. Here, we ask whether a machine-learning model can learn energetically relevant interaction preferences directly from native structures, without affinity or mutation labels. We decompose interfaces into local structural motifs and learn the compatibility of each motif with its surrounding molecular environment. Applied across a protein surface, these models yield target preference maps (TPMs), spatial fields of local interaction compatibility. Trained exclusively on native antibody-antigen structures without affinity or mutation labels, TPMs recover charged, aromatic and backbone-mediated recognition across peptide-protein and other non-antibody interfaces. On an out-of-domain SKEMPI benchmark, wild-type TPM scores discriminated mutation-sensitive positions (AUC 0.643, n=1323), while TPM-based substitution scores ranked alternative amino acids comparably to FoldX and Rosetta Flex {Delta}{Delta}G. These results show that native interface structures can supervise learning of transferable, energetically relevant molecular preferences without direct thermodynamic labels.
Related
- RPFace: From biased structural benchmarks to generalizable RNA-protein interface prediction
- Toward a Transferable Representation of Protein Dynamics for Functional Prediction and Mechanistic Discovery
- Microbial Protein as a novel ingredient in pet food: A food preference study in dogs and cats
- Foldseek-Interface reveals a protein interface universe far from complete
- Brain Landscape In Situ Crosslinking Mass Spectrometry (BLIS-XL-MS) Enables Global Analysis of Protein Structural Remodeling in the Brain
BibTeX
@misc{kienlein2026local,
title = {{Local structural preference maps encode transferable protein-interface energetics}},
author = {M. Kienlein and F. Menezes and P. Grass and S. Munker and T. Siebenmorgen and G. M. Popowicz},
howpublished = {bioRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.18.752156},
url = {https://www.biorxiv.org/content/10.64898/2026.09.18.752156v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Local structural preference maps encode transferable protein-interface energetics
AU - M. Kienlein
AU - F. Menezes
AU - P. Grass
AU - S. Munker
AU - T. Siebenmorgen
AU - G. M. Popowicz
PY - 2026
DA - 2026/09/18
PB - bioRxiv
DO - 10.64898/2026.09.18.752156
UR - https://www.biorxiv.org/content/10.64898/2026.09.18.752156v1
AB - Relating the drivers of binding affinity to structural features remains difficult because affinity emerges from many weak, context-dependent interactions, while experimental affinity and mutation data are limited. Here, we ask whether a machine-learning model can learn energetically relevant interaction preferences directly from native structures, without affinity or mutation labels. We decompose interfaces into local structural motifs and learn the compatibility of each motif with its surrounding molecular environment. Applied across a protein surface, these models yield target preference maps (TPMs), spatial fields of local interaction compatibility. Trained exclusively on native antibody-antigen structures without affinity or mutation labels, TPMs recover charged, aromatic and backbone-mediated recognition across peptide-protein and other non-antibody interfaces. On an out-of-domain SKEMPI benchmark, wild-type TPM scores discriminated mutation-sensitive positions (AUC 0.643, n=1323), while TPM-based substitution scores ranked alternative amino acids comparably to FoldX and Rosetta Flex {Delta}{Delta}G. These results show that native interface structures can supervise learning of transferable, energetically relevant molecular preferences without direct thermodynamic labels.
ER -