ToxinPred2: an improved method for predicting toxicity of proteins

Briefings in Bioinformatics, Journal Year: 2022, Volume and Issue: 23(5)

Published: April 20, 2022

Proteins/peptides have shown to be promising therapeutic agents for a variety of diseases. However, toxicity is one the obstacles in protein/peptide-based therapy. The current study describes web-based tool, ToxinPred2, developed predicting proteins. This an update ToxinPred mainly peptides and small method has been trained, tested evaluated on three datasets curated from recent release SwissProt. To provide unbiased evaluation, we performed internal validation 80% data external remaining 20% data. We implemented following techniques protein toxicity; (i) Basic Local Alignment Search Tool-based similarity, (ii) Motif-EmeRging with Classes-Identification-based motif search (iii) Prediction models. Similarity motif-based achieved high probability correct prediction poor sensitivity/coverage, whereas models based machine-learning balance sensitivity specificity reasonably accuracy. Finally, hybrid that combined all approaches maximum area under receiver operating characteristic curve around 0.99 Matthews correlation coefficient 0.91 dataset. In addition, alternate realistic datasets. best machine learning web server named 'ToxinPred2', which available at https://webs.iiitd.edu.in/raghava/toxinpred2/ standalone version https://github.com/raghavagps/toxinpred2. general proteins regardless their source origin.

Language: Английский

---

Machine learning for antimicrobial peptide identification and design

Nature Reviews Bioengineering, Journal Year: 2024, Volume and Issue: 2(5), P. 392 - 407

Published: Feb. 26, 2024

Language: Английский

---

ToxinPred 3.0: An improved method for predicting the toxicity of peptides

Computers in Biology and Medicine, Journal Year: 2024, Volume and Issue: 179, P. 108926 - 108926

Published: July 21, 2024

Language: Английский

---

AI Methods for Antimicrobial Peptides: Progress and Challenges

Microbial Biotechnology, Journal Year: 2025, Volume and Issue: 18(1)

Published: Jan. 1, 2025

ABSTRACT Antimicrobial peptides (AMPs) are promising candidates to combat multidrug‐resistant pathogens. However, the high cost of extensive wet‐lab screening has made AI methods for identifying and designing AMPs increasingly important, with machine learning (ML) techniques playing a crucial role. approaches have recently revolutionised this field by accelerating discovery new anti‐infective activity, particularly in preclinical mouse models. Initially, classical ML dominated field, but there been shift towards deep (DL) Despite significant contributions, existing reviews not thoroughly explored potential large language models (LLMs), graph neural networks (GNNs) structure‐guided AMP design. This review aims fill that gap providing comprehensive overview latest advancements, challenges opportunities using methods, particular emphasis on LLMs, GNNs We discuss limitations current highlight most relevant topics address coming years

Language: Английский

---

THRONE: A New Approach for Accurate Prediction of Human RNA N7-Methylguanosine Sites

Journal of Molecular Biology, Journal Year: 2022, Volume and Issue: 434(11), P. 167549 - 167549

Published: March 16, 2022

Language: Английский

---

CSM-Toxin: A Web-Server for Predicting Protein Toxicity

Pharmaceutics, Journal Year: 2023, Volume and Issue: 15(2), P. 431 - 431

Published: Jan. 28, 2023

Biologics are one of the most rapidly expanding classes therapeutics, but can be associated with a range toxic properties. In small-molecule drug development, early identification potential toxicity led to significant reduction in clinical trial failures, however we currently lack robust qualitative rules or predictive tools for peptide- and protein-based biologics. To address this, have manually curated largest set high-quality experimental data on peptide protein toxicities, developed CSM-Toxin, novel in-silico classifier, which relies solely primary sequence. Our approach encodes sequence information using deep learning natural languages model understand "biological" language, where residues treated as words sequences sentences. The CSM-Toxin was able accurately identify peptides proteins toxicity, achieving an MCC up 0.66 across both cross-validation multiple non-redundant blind tests, outperforming other methods highlighting generalisable performance our model. We strongly believe will serve valuable platform minimise biologic development pipeline. method is freely available easy-to-use webserver.

Language: Английский

---

ToxinPred 3.0: An improved method for predicting the toxicity of peptides

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2023, Volume and Issue: unknown

Published: Aug. 14, 2023

Abstract Toxicity emerges as a prominent challenge in the design of therapeutic peptides, causing failure numerous peptides during clinical trials. In 2013, our group developed ToxinPred, computational method that has been extensively adopted by scientific community for predicting peptide toxicity. this paper, we propose refined variant ToxinPred showcases improved reliability and accuracy Initially, used BLAST alignment-based toxicity prediction, yet coverage was limited. We motif-based approach with MERCI software to identify unique toxic patterns. Despite specificity gains, sensitivity compromised. alignment-free methods using machine/deep learning, achieving balance prediction. A deep learning model (ANN – LSTM fixed sequence length) one-hot encoding attained 0.93 AUROC 0.71 MCC on independent data. The machine (extra tree) compositional features achieved 0.95 0.78 MCC. Lastly, hybrid or ensemble combining two more models enhance performance. Hybrid approaches, including 0.98 0.81 Evaluation data demonstrated method’s superiority. To cater needs community, have standalone software, pip package web-based server ToxinPred3 ( https://github.com/raghavagps/toxinpred3 https://webs.iiitd.edu.in/raghava/toxinpred3/ ) . Author’s Biography Anand Singh Rathore is currently pursuing Ph.D. Computational Biology at Department Biology, Indraprastha Institute Information Technology, New Delhi, India. Akanksha Arora Shubham Choudhury Purava Tijare Project Fellow Gajendra P. S. Raghava working Professor Head Highlights Implementation alignment similarly based techniques peptides. Discovery toxicity-associated patterns identification regions Development learning-based Ensemble combine methods. Web screening peptides/proteins.

Language: Английский

---

The role and future prospects of artificial intelligence algorithms in peptide drug development

Biomedicine & Pharmacotherapy, Journal Year: 2024, Volume and Issue: 175, P. 116709 - 116709

Published: May 6, 2024

Peptide medications have been more well-known in recent years due to their many benefits, including low side effects, high biological activity, specificity, effectiveness, and so on. Over 100 peptide introduced the market treat a variety of illnesses. Most these are developed on basis endogenous peptides or natural peptides, which frequently required expensive, time-consuming, extensive tests confirm. As artificial intelligence advances quickly, it is now possible build machine learning deep models that screen large number candidate sequences for therapeutic peptides. Therapeutic such as those with antibacterial anticancer properties, by application algorithms.The process finding developing drugs outlined this review, along few related cases were helped AI conventional methods. These resources will open up new avenues drug development discovery, helping meet pressing needs clinical patients disease treatment. Although class biopharmaceuticals distinguish them from chemical small molecule drugs, purpose value cannot be ignored. However, traditional research has long cycle investment, creation substantially hastened AI-assisted (AI+) mode, offering boost combating diseases.

Language: Английский

---

De novo antioxidant peptide design via machine learning and DFT studies

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: March 18, 2024

Abstract Antioxidant peptides (AOPs) are highly valued in food and pharmaceutical industries due to their significant role human function. This study introduces a novel approach identifying robust AOPs using deep generative model based on sequence representation. Through filtration with deep-learning classification subsequent clustering via the Butina cluster algorithm, twelve ( GP1–GP12 ) potential antioxidant capacity were predicted. Density functional theory (DFT) calculations guided selection of six for synthesis biological experiments. Molecular orbital representations revealed that HOMO these is primarily localized indole segment, underscoring its pivotal activity. All synthesized exhibited activity DPPH assay, while hydroxyl radical test showed suboptimal results. A hemolysis assay confirmed non-hemolytic nature generated peptides. Additionally, an silico investigation explored inhibitory interaction between Keap1 protein. Analysis ligands GP3 , GP4 GP12 induced structural changes proteins, affecting stability flexibility. These findings highlight capability machine learning approaches generating

Language: Английский

---

Peptide nanozymes: An emerging direction for functional enzyme mimics

Bioactive Materials, Journal Year: 2024, Volume and Issue: 42, P. 284 - 298

Published: Sept. 4, 2024

Language: Английский

---