Integration of pre-trained protein language models with equivariant graph neural networks for peptide toxicity prediction

Abstract Background Peptide-based therapeutics have great potential due to their versatility, high specificity, and suitability for a variety of therapeutic applications. Despite these advantages, the inherent toxicities of some peptides pose challenges in drug development. Several computational met...

Full description

Saved in:
Bibliographic Details
Main Authors: Shihu Jiao, Xiucai Ye, Tetsuya Sakurai, Quan Zou, Wu Han, Chao Zhan
Format: Article
Language:English
Published: BMC 2025-07-01
Series:BMC Biology
Subjects:
Online Access:https://doi.org/10.1186/s12915-025-02329-1
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849331580653273088
author Shihu Jiao
Xiucai Ye
Tetsuya Sakurai
Quan Zou
Wu Han
Chao Zhan
author_facet Shihu Jiao
Xiucai Ye
Tetsuya Sakurai
Quan Zou
Wu Han
Chao Zhan
author_sort Shihu Jiao
collection DOAJ
description Abstract Background Peptide-based therapeutics have great potential due to their versatility, high specificity, and suitability for a variety of therapeutic applications. Despite these advantages, the inherent toxicities of some peptides pose challenges in drug development. Several computational methods have been developed to allow rapid and efficient large-scale screening of peptide toxicity. However, these methods mainly rely on the primary sequence and often ignore critical structural information, which limits their predictive accuracy. Results In this study, we introduce a novel framework named StrucToxNet that integrates a pre-trained protein language model with an equivariant graph neural network to improve peptide toxicity prediction. By combining sequence embeddings from the ProtT5 language model and 3D structural data predicted by ESMFold, StrucToxNet can capture both sequential and spatial characteristics of peptides. Testing on the independent dataset indicates that StrucToxNet outperforms existing sequence-based models in various metrics, achieving higher balanced accuracy and overall performance. Conclusions The results demonstrate the robustness and generalizability of StrucToxNet, marking it a reliable tool in the computational screening of toxic peptides and facilitating safer peptide-based drug development.
format Article
id doaj-art-dca9fc95c5514cf79d2002734e45bda2
institution Kabale University
issn 1741-7007
language English
publishDate 2025-07-01
publisher BMC
record_format Article
series BMC Biology
spelling doaj-art-dca9fc95c5514cf79d2002734e45bda22025-08-20T03:46:29ZengBMCBMC Biology1741-70072025-07-0123111010.1186/s12915-025-02329-1Integration of pre-trained protein language models with equivariant graph neural networks for peptide toxicity predictionShihu Jiao0Xiucai Ye1Tetsuya Sakurai2Quan Zou3Wu Han4Chao Zhan5Yangtze Delta Region Institute (Quzhou), University of Electronic Science and Technology of ChinaDepartment of Computer Science, University of TsukubaDepartment of Computer Science, University of TsukubaYangtze Delta Region Institute (Quzhou), University of Electronic Science and Technology of ChinaDepartment of Statistics, Stanford UniversityHarbin Medical University Cancer HospitalAbstract Background Peptide-based therapeutics have great potential due to their versatility, high specificity, and suitability for a variety of therapeutic applications. Despite these advantages, the inherent toxicities of some peptides pose challenges in drug development. Several computational methods have been developed to allow rapid and efficient large-scale screening of peptide toxicity. However, these methods mainly rely on the primary sequence and often ignore critical structural information, which limits their predictive accuracy. Results In this study, we introduce a novel framework named StrucToxNet that integrates a pre-trained protein language model with an equivariant graph neural network to improve peptide toxicity prediction. By combining sequence embeddings from the ProtT5 language model and 3D structural data predicted by ESMFold, StrucToxNet can capture both sequential and spatial characteristics of peptides. Testing on the independent dataset indicates that StrucToxNet outperforms existing sequence-based models in various metrics, achieving higher balanced accuracy and overall performance. Conclusions The results demonstrate the robustness and generalizability of StrucToxNet, marking it a reliable tool in the computational screening of toxic peptides and facilitating safer peptide-based drug development.https://doi.org/10.1186/s12915-025-02329-1PeptidesToxicityStructureProtein language modelEquivariant graph neural networks
spellingShingle Shihu Jiao
Xiucai Ye
Tetsuya Sakurai
Quan Zou
Wu Han
Chao Zhan
Integration of pre-trained protein language models with equivariant graph neural networks for peptide toxicity prediction
BMC Biology
Peptides
Toxicity
Structure
Protein language model
Equivariant graph neural networks
title Integration of pre-trained protein language models with equivariant graph neural networks for peptide toxicity prediction
title_full Integration of pre-trained protein language models with equivariant graph neural networks for peptide toxicity prediction
title_fullStr Integration of pre-trained protein language models with equivariant graph neural networks for peptide toxicity prediction
title_full_unstemmed Integration of pre-trained protein language models with equivariant graph neural networks for peptide toxicity prediction
title_short Integration of pre-trained protein language models with equivariant graph neural networks for peptide toxicity prediction
title_sort integration of pre trained protein language models with equivariant graph neural networks for peptide toxicity prediction
topic Peptides
Toxicity
Structure
Protein language model
Equivariant graph neural networks
url https://doi.org/10.1186/s12915-025-02329-1
work_keys_str_mv AT shihujiao integrationofpretrainedproteinlanguagemodelswithequivariantgraphneuralnetworksforpeptidetoxicityprediction
AT xiucaiye integrationofpretrainedproteinlanguagemodelswithequivariantgraphneuralnetworksforpeptidetoxicityprediction
AT tetsuyasakurai integrationofpretrainedproteinlanguagemodelswithequivariantgraphneuralnetworksforpeptidetoxicityprediction
AT quanzou integrationofpretrainedproteinlanguagemodelswithequivariantgraphneuralnetworksforpeptidetoxicityprediction
AT wuhan integrationofpretrainedproteinlanguagemodelswithequivariantgraphneuralnetworksforpeptidetoxicityprediction
AT chaozhan integrationofpretrainedproteinlanguagemodelswithequivariantgraphneuralnetworksforpeptidetoxicityprediction