Generative β-hairpin design using a residue-based physicochemical property landscape DOI Creative Commons
Vardhan Satalkar, Gemechis D. Degaga, Wei Li

и другие.

Biophysical Journal, Год журнала: 2024, Номер 123(17), С. 2790 - 2806

Опубликована: Фев. 1, 2024

De novo peptide design is a new frontier that has broad application potential in the biological and biomedical fields. Most existing models for de are largely based on sequence homology can be restricted evolutionarily derived protein sequences lack physicochemical context essential folding. Generative machine learning promising way to synthesize theoretical data on, but unique from, observable universe. In this study, we created tested custom generative adversarial network intended fold into β-hairpin secondary structure. This deep neural model designed establish preliminary foundation of approach conformational properties 20 canonical amino acids, example, hydrophobicity residue volume, using extant structure-specific from PDB. The beta robustly distinguishes structures β hairpin α helix intrinsically disordered peptides with an accuracy up 96% generates artificial minimum identities around 31% 50% when compared against current NCBI PDB nonredundant databases, respectively. These results highlight specifically anchored by property features acids expand sequence-to-structure landscape proteins beyond evolutionary limits.

Язык: Английский

DeepBindGCN: Integrating Molecular Vector Representation with Graph Convolutional Neural Networks for Protein–Ligand Interaction Prediction DOI Creative Commons
Haiping Zhang, Konda Mani Saravanan, John Z. H. Zhang

и другие.

Molecules, Год журнала: 2023, Номер 28(12), С. 4691 - 4691

Опубликована: Июнь 10, 2023

The core of large-scale drug virtual screening is to select the binders accurately and efficiently with high affinity from large libraries small molecules in which non-binders are usually dominant. binding significantly influenced by protein pocket, ligand spatial information, residue types/atom types. Here, we used pocket residues or atoms as nodes constructed edges neighboring information comprehensively represent information. Moreover, model pre-trained molecular vectors performed better than one-hot representation. main advantage DeepBindGCN that it independent docking conformation, concisely keeps physical-chemical features. Using TIPE3 PD-L1 dimer proof-of-concept examples, proposed a pipeline integrating other methods identify strong-binding-affinity compounds. It first time non-complex-dependent has achieved root mean square error (RMSE) value 1.4190 Pearson r 0.7584 PDBbind v.2016 set, respectively, thereby showing comparable prediction power state-of-the-art models rely upon 3D complex. provides powerful tool predict protein-ligand interaction can be many important application scenarios.

Язык: Английский

Процитировано

14
A deep learning based multi-model approach for predicting drug-like chemical compound’s toxicity DOI
Konda Mani Saravanan,

Jiang-Fan Wan,

Liujiang Dai

и другие.

Methods, Год журнала: 2024, Номер 226, С. 164 - 175

Опубликована: Май 1, 2024

Язык: Английский

Процитировано

5
Exploring the Modern Bioactive Peptides (BAPs) Universe: Doors to the Future DOI
Prasanna J. Patil

Peptide Science, Год журнала: 2024, Номер unknown

Опубликована: Июнь 15, 2024

ABSTRACT Because of their wide variety biological effects, bioactive peptides (BAPs) have recently attracted a lot attention. BAPs been observed to be safe, thanks widely acknowledged safety status by the United States Food and Drug Administration (USFDA). This has led widespread use in various industries, such as food nutrition, pharmaceuticals, therapeutics. A considerable amount research devoted developing cutting‐edge nanomaterials derived from BAPs, which utilized range industries. In realm scientific research, remarkable ability self‐assemble harnessed develop nanoassemblies. These nanoassemblies hold immense potential for advancement biomaterials future. Research interest continues focus on study detection using artificial intelligence (AI). Over past few years, there surge utilizing bio‐inspired strategies explore new possibilities development advanced energy devices storage solutions. However, these require extensive review offers broad perspective applications nanotechnology well pharmaceuticals Moreover, silico analysis coupled with ‐omics techniques, discussed. bargain, next‐generation approaches BAP comprising BAP‐based devices, AI, catalogued. There is emphasis more eco‐friendly energy‐storage technologies that draw inspiration nature BAPs.

Язык: Английский

Процитировано

5
BIBLIOMETRIC ANALYSIS OF ARTIFICIAL INTELLIGENCE IN HEALTHCARE RESEARCH: TRENDS AND FUTURE DIRECTIONS DOI Creative Commons
Renganathan Senthil, Thirunavukarasou Anand,

Chaitanya Sree Somala

и другие.

Future Healthcare Journal, Год журнала: 2024, Номер 11(3), С. 100182 - 100182

Опубликована: Сен. 1, 2024

The presence of artificial intelligence (AI) in healthcare is a powerful and game-changing force that completely transforming the industry as whole. Using sophisticated algorithms data analytics, AI has unparalleled prospects for improving patient care, streamlining operational efficiency, fostering innovation across ecosystem. This study conducts comprehensive bibliometric analysis research on healthcare, utilising SCOPUS database primary source.

Язык: Английский

Процитировано

5
Reinforcement learning-driven exploration of peptide space: accelerating generation of drug-like peptides DOI Creative Commons
Qian Wang,

Xiaotong Hu,

Zhiqiang Wei

и другие.

Briefings in Bioinformatics, Год журнала: 2024, Номер 25(5)

Опубликована: Июль 25, 2024

Abstract Using amino acid residues in peptide generation has solved several key problems, including precise control of sequence order, customized peptides for property modification, and large-scale synthesis. Proteins contain unknown residues. Extracting them the synthesis drug-like can create novel structures with unique properties, driving drug development. Computer-aided design molecules solve high-cost low-efficiency problems traditional discovery process. Previous studies faced limitations enhancing bioactivity drug-likeness polypeptide drugs due to less emphasis on connection relationships structures. Thus, we proposed a reinforcement learning-driven model based graph attention mechanisms generation. By harnessing advantages mechanisms, this effectively captured connectivity between peptides. Simultaneously, leveraging learning’s strength guiding optimal searches provided approach optimization. This introduces an actor-critic framework real-time feedback loops achieve dynamic balance attributes, which customize multiple specific targets enhance affinity targets. Experimental results demonstrate that generated meet specified absorption, distribution, metabolism, excretion, toxicity properties success rate over 90$\%$, thereby significantly accelerating process

Язык: Английский

Процитировано

4
Artificial intelligence for the discovery of antimicrobial peptides DOI
Paola Ruiz Puentes, Nicolás Aparicio, Pablo Arbeláez

и другие.

Elsevier eBooks, Год журнала: 2025, Номер unknown, С. 59 - 79

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

0
Single amino acid substitution analogs of marine antioxidant peptides with membrane permeability exert a marked protective effect against ultraviolet-B induced damage DOI
Yichao Huang, Qianjun He,

Pei-Pei Zhang

и другие.

Journal of Photochemistry and Photobiology B Biology, Год журнала: 2025, Номер unknown, С. 113120 - 113120

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

0
Advances of deep Neural Networks (DNNs) in the development of peptide drugs DOI
Yuzhen Niu,

Pingyang Qin,

Ping Lin

и другие.

Future Medicinal Chemistry, Год журнала: 2025, Номер unknown, С. 1 - 15

Опубликована: Фев. 12, 2025

Peptides are able to bind difficult disease targets with high potency and specificity, providing great opportunities meet unmet medical requirements. Nevertheless, the unique features of peptides, such as their small size, structural flexibility, scarce data availability, bring extra challenges design process. Firstly, this review sums up application peptide drugs in treating diseases. Then, probes into advantages Deep Neural Networks (DNNs) predicting designing structures. DNNs have demonstrated remarkable capabilities prediction, enabling accurate three-dimensional modeling through models like AlphaFold its successors. Finally, deliberates on coping strategies development drugs, along future research directions. Future directions focus further improving accuracy efficiency DNN-based drug design, exploring novel applications accelerating clinical translation. With continuous advancements technology accumulation, poised play an increasingly crucial role field development.

Язык: Английский

Процитировано

0
A fast and efficient virtual screening and identification strategy for helix peptide binders based on finDr webserver: A case study of bovine serum albumin (BSA) DOI

J. Bu,

Na Luo, Cheng Shen

и другие.

International Journal of Biological Macromolecules, Год журнала: 2025, Номер unknown, С. 141118 - 141118

Опубликована: Фев. 1, 2025

Язык: Английский

Процитировано

0
Molecular Modelling in Bioactive Peptide Discovery and Characterisation DOI Creative Commons
Clement Agoni, Raúl Fernández-Díaz, Patrick Brendan Timmons

и другие.

Biomolecules, Год журнала: 2025, Номер 15(4), С. 524 - 524

Опубликована: Апрель 3, 2025

Molecular modelling is a vital tool in the discovery and characterisation of bioactive peptides, providing insights into their structural properties interactions with biological targets. Many models predicting peptide function or structure rely on intrinsic properties, including influence amino acid composition, sequence, chain length, which impact stability, folding, aggregation, target interaction. Homology predicts structures based known templates. Peptide–protein can be explored using molecular docking techniques, but there are challenges related to inherent flexibility addressed by more computationally intensive approaches that consider movement over time, called dynamics (MD). Virtual screening many usually against single target, enables rapid identification potential peptides from large libraries, typically approaches. The integration artificial intelligence (AI) has transformed leveraging amounts data. AlphaFold general protein prediction deep learning greatly improved predictions conformations interactions, addition estimates model accuracy at each residue guide interpretation. Peptide being further enhanced Protein Language Models (PLMs), deep-learning-derived statistical learn computer representations useful identify fundamental patterns proteins. Recent methodological developments discussed context canonical as well those modifications cyclisations. In designing therapeutics, main outstanding challenge for these methods incorporation diverse non-canonical acids

Язык: Английский

Процитировано

0