A Systematic Benchmark for Peptide Property Prediction
Xiaoying Dong , Kaijun Yang , Tianxiang Wu , Pengyong Li , Lin Gao
Accurate prediction of peptide physicochemical properties and biological activities is critical for rational peptide design. However, current research is often constrained by heterogeneous data sources and inconsistent evaluation standards, which hinder fair comparisons and reliable assessments of model generalization. We present a peptide property prediction benchmark designed to evaluate model performance with an emphasis on realistic generalization across both classification and regression tasks. By applying unified biological filtering criteria, we systematically curated and standardized 15 datasets comprising 161 571 sequences, spanning a wide range of physicochemical properties and functional activities. We benchmarked eight representative architectures—encompassing traditional machine learning, deep learning, and pretrained language models—alongside diverse feature encoding schemes. Furthermore, we investigated the impact of random versus similarity-based data splitting strategies on model robustness. Our benchmarks reveal that pretrained language models generally achieve strong overall performance, while the similarity-based evaluation exposes the limitations of current clustering algorithms on peptide data. Ultimately, this benchmark lays the foundation for developing next-generation predictive tools that integrate sequence information with structural and physicochemical context.
Peptidepropertyprediction / Benchmarkdataset / Modelevaluation
Higher Education Press 2026
/
| 〈 |
|
〉 |
PDF
(4225KB)
