Deep Learning-Based Bioactive Therapeutic Peptide Generation and Screening

Abstract

Many bioactive peptides demonstrated therapeutic effects over complicated diseases, such as antiviral, antibacterial, anticancer, etc. It is possible to generate a large number of potentially bioactive peptides using deep learning in a manner analogous to the generation of de novo chemical compounds using the acquired bioactive peptides as a training set. Such generative techniques would be significant for drug development since peptides are much easier and cheaper to synthesize than compounds. Despite the limited availability of deep learning-based peptide-generating models, we have built an LSTM model (called LSTM_Pep) to generate de novo peptides and fine-tuned the model to generate de novo peptides with specific prospective therapeutic benefits. Remarkably, the Antimicrobial Peptide Database has been effectively utilized to generate various kinds of potential active de novo peptides. We proposed a pipeline for screening those generated peptides for a given target and used the main protease of SARS-COV-2 as a proof-of-concept. Moreover, we have developed a deep learning-based protein–peptide prediction model (DeepPep) for rapid screening of the generated peptides for the given targets. Together with the generating model, we have demonstrated that iteratively fine-tuning training, generating, and screening peptides for higher-predicted binding affinity peptides can be achieved. Our work sheds light on developing deep learning-based methods and pipelines to effectively generate and obtain bioactive peptides with a specific therapeutic effect and showcases how artificial intelligence can help discover de novo bioactive peptides that can bind to a particular target.

Therapeutic peptides for coronary artery diseases: in silico methods and current perspectives

“…Plisson et al, applied several machine learning algorithms to hemolytic peptide and antimicrobial peptide datasets and emphasized that gradient boosting and extreme gradient boosting classifiers performed the best and even proposed 34 high-confidence nonhemolytic natural AMPs (Plisson et al 2020 ). A more recent approach has shown that by enhancing the effectiveness of machine learning methods with molecular docking and dynamics simulations, it is possible to develop fine-tuned peptide designs that can achieve the desired biological activity (Zhang et al 2023 ). In particular, the use of machine/deep learning methods, which are frequently used in the design of anticancer and microbial peptides and in the study of protein‒peptide interactions, in the design of novel therapeutic molecules for CAD could lead to significant advances (Varadi et al 2022 ; Jumper et al 2021 ; Lei et al 2021 ).…”

Exploring the Potential of Bioactive Peptides: From Natural Sources to Therapeutics

“…Numerous data-driven computational methods, particularly artificial intelligence (AI)-based support vector machine (SVM), random forest (RF), extremely randomized trees (ERTs), and deep learning (DL), have been developed to assist in predicting a large pool of therapeutic peptides based on peptide datasets generated by high-performance sequencing [138,139]. For instance, deep learning models for peptide generation such as DeepImmuno-GAN (architecture developed for generating potential MHC-binding peptides) [140], DeepACP and XDeep-AcPEP (for identifying and predicting activities of anticancer peptides) [141,142], ProteinGAN [143], HydrAMP [144], peptide VAE [145], pepGAN [146], and pepVAE [147] have been developed [148].…”

A Multi-Modal Contrastive Diffusion Model for Therapeutic Peptide Generation

“…Datasets. Following previous studies (Thi Phan et al 2022;Zhang et al 2023a), we collected therapeutic peptide data from public databases, containing two biological types, i.e., antimicrobial peptides (AMP) and anticancer peptides (ACP). Among these collected peptides, a portion of them only have 1D sequence information, without 3D structure information.…”