Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Hsu, Wei-Tse, Grevtsev, Savva, Douglas, Thomas, Magarkar, Aniket, Biggin, Philip C.
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2507.07882
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866918088351416320
author Hsu, Wei-Tse
Grevtsev, Savva
Douglas, Thomas
Magarkar, Aniket
Biggin, Philip C.
author_facet Hsu, Wei-Tse
Grevtsev, Savva
Douglas, Thomas
Magarkar, Aniket
Biggin, Philip C.
contents We evaluate the feasibility of using co-folding models for synthetic data augmentation in training machine learning-based scoring functions (MLSFs) for binding affinity prediction. Our results show that performance gains depend critically on the structural quality of augmented data. In light of this, we established simple heuristics for identifying high-quality co-folding predictions without reference structures, enabling them to substitute for experimental structures in MLSF training. Our study informs future data augmentation strategies based on co-folding models.
format Preprint
id arxiv_https___arxiv_org_abs_2507_07882
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Can AI-predicted complexes teach machine learning to compute drug binding affinity?
Hsu, Wei-Tse
Grevtsev, Savva
Douglas, Thomas
Magarkar, Aniket
Biggin, Philip C.
Machine Learning
We evaluate the feasibility of using co-folding models for synthetic data augmentation in training machine learning-based scoring functions (MLSFs) for binding affinity prediction. Our results show that performance gains depend critically on the structural quality of augmented data. In light of this, we established simple heuristics for identifying high-quality co-folding predictions without reference structures, enabling them to substitute for experimental structures in MLSF training. Our study informs future data augmentation strategies based on co-folding models.
title Can AI-predicted complexes teach machine learning to compute drug binding affinity?
topic Machine Learning
url https://arxiv.org/abs/2507.07882