TY - JOUR AU - Breimann, Stephan AU - Kamp, Frits AU - Basset, Gabriele AU - Abou-Ajram, Claudia AU - Güner, Gökhan AU - Yanagida, Kanta AU - Okochi, Masayasu AU - Müller, Stephan A AU - Lichtenthaler, Stefan F AU - Langosch, Dieter AU - Frishman, Dmitrij AU - Steiner, Harald TI - Charting γ-secretase substrates by explainable AI. JO - Nature Communications VL - 16 IS - 1 SN - 2041-1723 CY - [London] PB - Springer Nature M1 - DZNE-2025-00769 SP - 5428 PY - 2025 AB - Proteases recognize substrates by decoding sequence information-an essential cellular process elusive when recognition motifs are absent. Here, we unravel this problem for γ-secretase, an intramembrane-cleaving protease associated with Alzheimer's disease and cancer, by developing Comparative Physicochemical Profiling (CPP), a sequence-based algorithm for identifying interpretable physicochemical features. We show that CPP deciphers a γ-secretase substrate signature with single-residue resolution, which can explain the conformational transitions observed in substrates upon γ-secretase binding. Using machine learning, we predict the entire human γ-secretase substrate scope, revealing numerous previously unknown substrates. Our approach outperforms state-of-the-art protein language models, improving prediction accuracy from 60 KW - Amyloid Precursor Protein Secretases: metabolism KW - Amyloid Precursor Protein Secretases: chemistry KW - Amyloid Precursor Protein Secretases: genetics KW - Humans KW - Machine Learning KW - Substrate Specificity KW - Alzheimer Disease: metabolism KW - Algorithms KW - Protein Binding KW - Amyloid Precursor Protein Secretases (NLM Chemicals) LB - PUB:(DE-HGF)16 C6 - pmid:40593564 C2 - pmc:PMC12219630 DO - DOI:10.1038/s41467-025-60638-z UR - https://pub.dzne.de/record/279438 ER -