{"database":"biostudies-literature","file_versions":[],"scores":null,"additional":{"submitter":["Li X"],"funding":["Chongqing Talent Program Project","the Technology Innovation and Application Development Key Project of Chongqing","the National Natural Science Foundation of China"],"pagination":["4143"],"full_dataset_link":["https://www.ebi.ac.uk/biostudies/studies/S-EPMC12692575"],"repository":["biostudies-literature"],"omics_type":["Unknown"],"volume":["14(23)"],"pubmed_abstract":["Accurate prediction of hemolytic peptides is essential for peptide safety evaluation and therapeutic design; however, existing models remain constrained by limited accuracy and interpretability. To overcome these challenges, we propose a regression framework that integrates embeddings from a protein language model with handcrafted amino acid descriptors. Specifically, sequence representations derived from the ESM2_t33 model are fused with physicochemical amino acid descriptor features, and key predictive variables are selected through a three-stage strategy involving variance filtering, F-test ranking, and mutual information analysis. The final ensemble model, composed of Random Forest, Extremely Randomized Trees, Gradient Boosting, eXtreme Gradient Boosting (XGBoost), and Ridge Regression, achieved a coefficient of determination (R<sup>2</sup>) of 0.57 and a correlation coefficient (R) of 0.76 on the test set, outperforming previous approaches. To enhance interpretability, we applied Shapley value analysis and the Calibrated_Explanation algorithm to quantify feature contributions and generate reliable sample-specific explanations. The trained model has been deployed online as HemPepPred, a tool for predicting hemolytic concentration (HC<sub>50</sub>) values, which provides a practical platform for rational peptide design and safety assessment."],"journal":["Foods (Basel, Switzerland)"],"pubmed_title":["HemPepPred: Quantitative Prediction of Peptide Hemolytic Activity Based on Machine Learning and Protein Language Model-Derived Features."],"pmcid":["PMC12692575"],"funding_grant_id":["cstc2024ycjhbgzxm0113","CSTB2024TIAD-KPX0014","32172196"],"pubmed_authors":["Li X","Zhuo X","Liang G","Liang X","Yu S","Zhao W"],"additional_accession":[]},"is_claimable":false,"name":"HemPepPred: Quantitative Prediction of Peptide Hemolytic Activity Based on Machine Learning and Protein Language Model-Derived Features.","description":"Accurate prediction of hemolytic peptides is essential for peptide safety evaluation and therapeutic design; however, existing models remain constrained by limited accuracy and interpretability. To overcome these challenges, we propose a regression framework that integrates embeddings from a protein language model with handcrafted amino acid descriptors. Specifically, sequence representations derived from the ESM2_t33 model are fused with physicochemical amino acid descriptor features, and key predictive variables are selected through a three-stage strategy involving variance filtering, F-test ranking, and mutual information analysis. The final ensemble model, composed of Random Forest, Extremely Randomized Trees, Gradient Boosting, eXtreme Gradient Boosting (XGBoost), and Ridge Regression, achieved a coefficient of determination (R<sup>2</sup>) of 0.57 and a correlation coefficient (R) of 0.76 on the test set, outperforming previous approaches. To enhance interpretability, we applied Shapley value analysis and the Calibrated_Explanation algorithm to quantify feature contributions and generate reliable sample-specific explanations. The trained model has been deployed online as HemPepPred, a tool for predicting hemolytic concentration (HC<sub>50</sub>) values, which provides a practical platform for rational peptide design and safety assessment.","dates":{"release":"2025-01-01T00:00:00Z","publication":"2025 Dec","modification":"2026-05-26T14:07:08.214Z","creation":"2026-05-24T03:11:07.81Z"},"accession":"S-EPMC12692575","cross_references":{"pubmed":["41376080"],"doi":["10.3390/foods14234143"]}}