AI Insight
This study benchmarked machine learning methods for predicting peptide-protein binding affinity using 11,349 binding measurements across three different data partitioning strategies. The authors found that model performance varied substantially depending on whether test data contained similar peptides, the same protein targets, or completely new targets, with Spearman correlations of 0.462, 0.669, and 0.530 respectively. The choice of molecular representation had a greater impact on performance than the choice of regression algorithm, and the best-performing representation changed depending on the partitioning strategy.
Why it matters
Accurate prediction of peptide-protein binding is crucial for drug discovery and understanding biological interactions. This work demonstrates that evaluation methods significantly affect apparent model performance, highlighting the need for researchers to choose benchmarking strategies that match their intended application and to test models under realistic conditions where target proteins may be completely novel.
Understand the Science
⚠️ Preprint – Noch nicht peer-reviewed
Dieser Artikel wurde noch nicht von unabhängigen Experten begutachtet. Die Ergebnisse sind vorläufig und sollten mit Vorsicht interpretiert werden.
Abstract: Peptide-protein affinity models are often evaluated with a single data split, obscuring whether they interpolate among measurements for observed targets or generalize across peptide or target shifts. We integrated three sources of quantitative peptide-protein binding data to obtain 11,349 deduplicated pairs and benchmarked ten peptide representations, ESM-2 protein embeddings, and six regressors under peptide-similarity, within-target, and leave-target-out partitions. Across 60 matched representation-regressor configurations, mean test Spearman correlations were 0.462, 0.669, and 0.530, respectively. The top configuration shifted from ECFP-16 count fingerprints with random forest in the first two settings to HELM-BERT with Extra Trees when exact target sequences were excluded. Representation-rank correlations ranged from -0.042 to 0.624 across partitions, whereas regressor-rank correlations ranged from 0.771 to 0.943. Learning curves showed that representation differences were largest with limited supervision and narrowed as training data increased. PeptideCLM-2 adaptation and simple element-wise interaction features provided no consistent gain over a frozen encoder and direct concatenation under the tested protocols. These conclusions are specific to a dataset that pools transformed Kd, Ki, and IC50 measurements and to target exclusion at the exact-sequence level. Peptide-protein affinity benchmarks should therefore align data partitions with the intended use and jointly assess the effects of data scale, molecular representation, and downstream learner.
Source: Benchmarking Peptide-Protein Affinity Prediction Across Peptide and Target Shifts