GM-Pep: A High Efficiency Strategy to De Novo Design Functional Peptide Sequences

Abstract

Although peptides are regarded as ideal therapeutic agents, only a small proportion of the marketed drugs are peptides. In the past decade, pharmacists have paid great attention to the development of peptide therapeutics. Except a few approved chemically/rationally designed peptides, most attempts failed due to unsatisfactory efficacy or safety. Luckily, computation methods, such as artificial intelligence, have been utilized to accelerate the discovery of therapeutic peptides by predicting the activity, toxicity, and absorption, distribution, metabolism, and excretion of polypeptides. Usually, a specific biological activity of a peptide could be accurately determined by an interest-oriented binary classification constructed of a positive set and another unexperimentally validated negative set regardless of other characteristics, which suggests that it could be challenging to realize the comprehensive evaluation of the research object in the early stage of drug research and development. Herein, we proposed an integrated method (GM-Pep) that contained a conditional variational autoencoder model (CVAE) and a positive sample training multiclassifier (Deep-Multiclassifier) to effectively generate a single bioactive peptide sequence without toxicity and referential side effects. The results showed that our Deep-Multiclassifier model gave a sequence accuracy of up to 96.41% [toxicity (94.48%), antifungal (96.58%), antihypertensive (97.18%), and antibacterial (96.91%), respectively]. The properties of Deep-Multiclassifier and CVAE were validated through 12 first synthesized antibacterial peptides or compared to random peptides. The source code and data sets are available at https://github.com/TimothyChen225/GM-Pep.

Multifunctional Analysis of Chia Seed (Salvia hispanica L.) Bioactive Peptides Using Peptidomics and Molecular Dynamics Simulations Approaches

“…For example, generating libraries of large peptides is currently feasible, due to massive mass spectrometry techniques and/or peptidomics approaches. However, the prediction of properties, such as those related to the bioactivity of these biomolecules enables either simplifying assumptions to be made or the application of new approaches for predicting bioactivity [13]. Here, we conducted an in silico prediction of physicochemical properties (hydrophobicity, hydrophilicity, intestinal stability, antiangiogenic, antihypertensive, and anti-inflammatory, using molecular dynamics simulations (MDS) and ensemble docking-virtual screening (VS) methods, with potential bioactive peptides from chia seed, and correlated these properties with human molecular (protein) targets (angiotensin converting enzyme (ACE), vascular endothelial growth factor (VEGF), glucocorticoid (GLUC), and mineralocorticoid (MINC) receptors).…”

The discovery of antimicrobial peptides from the gut microbiome of cockroachBlattella germanicausing deep learning pipeline

“…The rational design requirements for AMPs have not yet been satisfied, and novel AMP sequences are mainly derived from high-throughput space-searching 45 . Most existing prediction models performed poorly in identifying non-AMP sequences, with only a 20-40% overall precision rate, indicating high false-positive ratios in real experiments [35][36][37][38][39][40][41][42][43][44]46 . Recently, a unified pipeline, incorporating 5 different complexing frameworks, was proposed to identify AMP from human gut microbiome data, and reached reasonable overall precision and specificities 46 .…”

“…In recent years, the identifications of AMPs took advantage of machine learning (ML) and Deep Learning (DL) [35][36][37] . These tools, whose AMPs prediction accuracy ranged from 87.17% to 92.11%, included support vector machine (SVM) 38 , k-nearest neighbor (KNN), 35 random forests (RFs) 39 , eXtreme Gradient Boosting (XGBoost) 40 , deep neural network (DNN) 41 , recurrent neural network (RNN) with long short-term memory (LSTM) 42 , etc. [43][44] .…”

Peptide-based drug discovery through artificial intelligence: towards an autonomous design of therapeutic peptides

“…Different VAE strategies have been implemented to assist the peptide design or discovery. Methods like PepVAE [ 107 ], PepCVAE [ 108 ] and GM-Pep [ 109 ] have been proposed for AMP design supported by VAE approaches. Two relevant methods for AMP design are proposed in [ 110 ] and [ 111 ].…”