Rerouting therapeutic peptides and unlocking their potential against SARS-CoV2

3 Biotech, Journal Year: 2025, Volume and Issue: 15(5)

Published: April 4, 2025

Language: Английский

---

Predicting host-pathogen interactions with machine learning algorithms: A scoping review

Infection Genetics and Evolution, Journal Year: 2025, Volume and Issue: 130, P. 105751 - 105751

Published: April 10, 2025

Language: Английский

---

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

Amino Acids, Journal Year: 2024, Volume and Issue: 56(1)

Published: May 31, 2024

Many drug formulations containing small active molecules are used for the treatment of coronary artery disease, which affects a significant part world's population. However, inadequate profile these in terms therapeutic efficacy has led to use protein and peptide-based biomolecules with superior properties, such as target-specific affinity low immunogenicity, critical diseases. Protein‒protein interactions, consequence advances molecular techniques strategies involving combined silico methods, have enabled design peptides reach an advanced dimension. In particular, advantages provided by protein/peptide structural modeling, docking study their dynamics simulations interactions under physiological conditions machine learning that can work combination all these, progress been made approaches developing modulate development progression this scope, review discusses methods peptide therapeutics disease identifying mechanisms be modulated designs provides comprehensive perspective future studies.

Language: Английский

---

Explainable Machine Learning Model to Accurately Predict Protein-Binding Peptides

Algorithms, Journal Year: 2024, Volume and Issue: 17(9), P. 409 - 409

Published: Sept. 12, 2024

Enzymes play key roles in the biological functions of living organisms, which serve as catalysts to and regulate biochemical reaction pathways. Recent studies suggest that peptides are promising molecules for modulating enzyme function due their advantages large chemical diversity well-established methods library synthesis. Experimental approaches identify protein-binding time-consuming costly. Hence, there is a demand develop fast accurate computational approach tackle this problem. Another challenge developing lack reliable dataset. In study, we new machine learning called PepBind-SVM predict peptides. To build model, extract different sequential physicochemical features from use Support Vector Machine (SVM) classification technique. We train model on dataset also introduce study. achieves 92.1% prediction accuracy, outperforming other classifiers at predicting

Language: Английский

---

Understanding and Therapeutic Application of Immune Response in Major Histocompatibility Complex (MHC) Diversity Using Multimodal Artificial Intelligence

BioMedInformatics, Journal Year: 2024, Volume and Issue: 4(3), P. 1835 - 1864

Published: Aug. 5, 2024

Human Leukocyte Antigen (HLA) is like a device that monitors the internal environment of body. T lymphocytes immediately recognize HLA molecules are expressed on surface cells different individual, attacking it defeats microorganisms one causes rejection in organ transplants performed between people with unmatched types. Over 2850 and 3580 polymorphisms have been reported for HLA-A HLA-B respectively, around world. genes associated risk developing variety diseases, including autoimmune play an important role pathological conditions. By using deep learning method called multi-task to simultaneously predict gene sequences multiple genes, possible improve accuracy shorten execution time. Some new systems use model convolutional neural network (CNNs) learning, which uses networks consisting many layers can learn complex correlations SNP information based reference data imputation, serves as training data. The learned output predicted values high input. To investigate part input surrounding used make predictions, predictions were made not only small number nearby but also distributed over wider area by visualizing model. While conventional methods strong at nearly good located distant locations, some thought prediction may improved because this problem was overcome. involved onset diseases attracting attention. As from perspective elucidating conditions realizing personalized medicine. applied two imputation panels—a Japanese panel (n = 1118) type I diabetes genetics consortium 5122). Through 10-fold cross-validation these panels, achieved higher than methods, especially imputing low-frequency rare alleles. increased expected increase reliability analysis, integrated analysis racial populations, greatly contribute identification further elucidation

Language: Английский

---

A quantitative intracellular peptide binding assay reveals recognition determinants and context dependence of short linear motifs

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 1, 2024

ABSTRACT Transient protein-protein interactions play key roles in controlling dynamic cellular responses. Many examples involve globular protein domains that bind to peptide sequences known as Short Linear Motifs (SLiMs), which are enriched intrinsically disordered regions of proteins. Here we describe a novel functional assay for measuring SLiM binding, called Systematic Intracellular Motif Binding Analysis (SIMBA). In this method, binding foreign domain its cognate allows yeast cells proliferate by blocking growth arrest signal. A high-throughput application the SIMBA method involving competitive and deep sequencing provides rapid quantification relative strength thousands sequence variants, comprehensive interrogation features control their recognition potency. We show multiple distinct classes SLiM-binding can be analyzed peptides vivo correlates with biochemical affinities measured vitro. Deep mutational scanning high-resolution definitions motif determinants reveals how variations at non-core positions modulate strength. Furthermore, parent human tankyrase ARC or YAP WW identifies modes uncovers context effects preferred residues one position depend on elsewhere. The findings establish fast incisive approach interrogating via massively parallel protein-peptide vivo.

Language: Английский

---

TPepPro: a deep learning model for predicting peptide-protein interactions

Bioinformatics, Journal Year: 2024, Volume and Issue: 41(1)

Published: Nov. 25, 2024

Abstract Motivation Peptides and their derivatives hold potential as therapeutic agents. The rising interest in developing peptide drugs is evidenced by increasing approval rates the FDA of USA. To identify most peptides, study on peptide-protein interactions (PepPIs) presents a very important approach but poses considerable technical challenges. In experimental aspects, transient nature PepPIs high flexibility peptides contribute to elevated costs inefficiency. Traditional docking molecular dynamics simulation methods require substantial computational resources, predictive accuracy results remain unsatisfactory. Results address this gap, we proposed TPepPro, Transformer-based model for PepPI prediction. We trained TPepPro dataset 19,187 pairs complexes with both sequential structural features. utilizes strategy that combines local protein sequence feature extraction global structure extraction. Moreover, optimizes architecture featuring neural network BN-ReLU arrangement, which notably reduced amount computing resources required According comparison analysis, reached 0.855 achieving an 8.1% improvement compared second-best TAGPPI. achieved AUC 0.922, surpassing TAGPPI 0.844. newly developed certain can be validated according previous evidence, thus indicating efficiency detect would helpful amino acid drug applications. Availability implementation source code available at https://github.com/wanglabhku/TPepPro.

Language: Английский

---

Leveraging Machine Learning to Enhance Information Exploration

SSRN Electronic Journal, Journal Year: 2024, Volume and Issue: unknown

Published: Jan. 1, 2024

Language: Английский

---

Leveraging Machine Learning to Enhance Information Exploration

Machine Learning and Applications An International Journal, Journal Year: 2024, Volume and Issue: 11(2), P. 17 - 27

Published: June 28, 2024

Machine learning algorithms are revolutionizing intelligent search and information discovery capabilities. By incorporating techniques like supervised learning, unsupervised reinforcement deep systems can automatically extract insights patterns from vast data repositories. Natural language processing enables deeper comprehension of text, while image recognition unlocks knowledge visual data. powers personalized recommendation engines accurate sentiment analysis. Integrating graphs enriches machine models with background for enhanced accuracy explainability. Applications span voice search, anomaly detection, predictive analytics, text mining, clustering. However, interpretable AI crucial enabling transparency trustworthiness. Key challenges include limited training data, complex domain requirements, ethical considerations around bias privacy. Ongoing research that combines representation, human-centered design will advance discovery. The collaboration between artificial human intelligence holds the potential to revolutionize access acquisition.

Language: Английский

---

In Silico Design of Peptide Inhibitors Targeting HER2 for Lung Cancer Therapy

Cancers, Journal Year: 2024, Volume and Issue: 16(23), P. 3979 - 3979

Published: Nov. 27, 2024

Background/Objectives: Human epidermal growth factor receptor 2 (HER2) is overexpressed in several malignancies, such as breast, gastric, ovarian, and lung cancers, where it promotes aggressive tumor proliferation unfavorable prognosis. Targeting HER2 has thus emerged a crucial therapeutic strategy, particularly for HER2-positive malignancies. The present study focusses on the design optimization of peptide inhibitors targeting HER2, utilizing machine learning to identify enhance candidates with elevated binding affinities. aim provide novel options malignancies linked overexpression. Methods: This started extraction structural examination protein, succeeded by designing sequences derived from essential interaction residues. A technique (XGBRegressor model) was employed predict affinities, identifying top 20 possibilities. underwent further screening via FreeSASA methodology free energy calculations, resulting selection four primary (pep-17, pep-7, pep-2, pep-15). Density functional theory (DFT) calculations were utilized evaluate molecular reactivity characteristics, while dynamics simulations performed investigate inhibitory mechanisms selectivity effects. Advanced computational methods, QM/MM simulations, offered more understanding peptide–protein interactions. Results: Among principal peptides, pep-7 exhibited most DFT values (−3386.93 kcal/mol) maximum dipole moment (10,761.58 Debye), whereas pep-17 had lowest value (−5788.49 minimal (2654.25 Debye). Molecular indicated that steady −12.88 kcal/mol consistently bound inside pocket during 300 ns simulation. showed overall total system, which combines both QM MM contributions, remained around −79,000 ± 400 kcal/mol, suggesting entire protein–peptide complex stable state, maintaining strong, well-integrated binding. Conclusions: Pep-7 promising peptide, displaying strong stability, favorable energy, stability HER2-overexpressing cancer models. These findings suggest viable candidate offering potential treatment strategy against HER2-driven

Language: Английский

---