Virtual screening of drugs targeting PD-L1 protein DOI Open Access

Kai-Dong Lin,

Xiaoqian Lin, Xubo Lin

et al.

Acta Physica Sinica, Journal Year: 2023, Volume and Issue: 72(24), P. 240501 - 240501

Published: Jan. 1, 2023

Monoclonal antibody inhibitors targeting PD-1/PD-L1 immune checkpoints are gradually entering the market and have achieved certain positive effects in treatments of various types tumors. However, with expansion application, limitations drugs problems such as excessive homogenization research appear, making small-molecule new focus researchers. This study aims to use ligand-based structure-based binding activity prediction methods achieve virtual screening PD-L1, thereby helping accelerate development small molecule drugs. A dataset PD-L1 inhibitory from relevant literature patents is collected judgment classification models intensity regression constructed based on different molecular featurization machine learning algorithms. The two filter 68 candidate compounds high a large drug-like pool (ZINC15). Ten these not only good drug similarities pharmacokinetics, but also exhibit comparable affinities similar mechanisms action previous reported hotspot docking. phenomenon further verified subsequent dynamics simulation estimation free energy. In this study, workflow integrating method developed, potential effectively screened compound databases, which expected help application tumor immunotherapy.

Language: Английский

De novo molecular design with deep molecular generative models for PPI inhibitors DOI
Jianmin Wang, Yanyi Chu, Jiashun Mao

et al.

Briefings in Bioinformatics, Journal Year: 2022, Volume and Issue: 23(4)

Published: July 13, 2022

We construct a protein-protein interaction (PPI) targeted drug-likeness dataset and propose deep molecular generative framework to generate novel molecules from the features of seed compounds. This gains inspiration published models, uses key associated with PPI inhibitors as input develops models for de novo design inhibitors. For first time, quantitative estimation index compounds targeting was applied evaluation generation model PPI-targeted Our results estimated that generated had better drug-likeness. Additionally, our also exhibits comparable performance other several state-of-the-art molecule models. The share chemical space iPPI-DB demonstrated by analysis. peptide characterization-oriented ligand-based are explored. Finally, we recommend this will be an important step forward therapeutics.

Language: Английский

Citations

49

Transformer-Based Molecular Generative Model for Antiviral Drug Design DOI Creative Commons
Jiashun Mao, Jianmin Wang, Amir Zeb

et al.

Journal of Chemical Information and Modeling, Journal Year: 2023, Volume and Issue: 64(7), P. 2733 - 2745

Published: June 27, 2023

Since the Simplified Molecular Input Line Entry System (SMILES) is oriented to atomic-level representation of molecules and not friendly in terms human readability editable, however, IUPAC closest natural language very human-oriented performing molecular editing, we can manipulate generate corresponding new produce programming-friendly forms SMILES. In addition, antiviral drug design, especially analogue-based also more appropriate edit design directly from functional group level than atomic SMILES, since designing analogues involves altering R only, which closer knowledge-based a chemist. Herein, present novel data-driven self-supervised pretraining generative model called "TransAntivirus" make select-and-replace edits convert organic into desired properties for candidate analogues. The results indicated that TransAntivirus significantly superior control models novelty, validity, uniqueness, diversity. showed excellent performance optimization nucleoside non-nucleoside by chemical space analysis property prediction analysis. Furthermore, validate applicability drugs, conducted two case studies on screened four lead compounds against anticoronavirus disease (COVID-19). Finally, recommend this framework accelerating discovery.

Language: Английский

Citations

37

Explore drug-like space with deep generative models DOI
Jianmin Wang, Jiashun Mao, Meng Wang

et al.

Methods, Journal Year: 2023, Volume and Issue: 210, P. 52 - 59

Published: Jan. 19, 2023

Language: Английский

Citations

23

Unlocking the Potential of Generative Artificial Intelligence in Drug Discovery DOI
Virgilio Romanelli, Carmen Cerchia, Antonio Lavecchia

et al.

Springer eBooks, Journal Year: 2024, Volume and Issue: unknown, P. 37 - 63

Published: Jan. 1, 2024

Language: Английский

Citations

9

Design of Cyclic Peptides Targeting Protein–Protein Interactions Using AlphaFold DOI Open Access
Takatsugu Kosugi, Masahito Ohue

International Journal of Molecular Sciences, Journal Year: 2023, Volume and Issue: 24(17), P. 13257 - 13257

Published: Aug. 26, 2023

More than 930,000 protein-protein interactions (PPIs) have been identified in recent years, but their physicochemical properties differ from conventional drug targets, complicating the use of small molecules as modalities. Cyclic peptides are a promising modality for targeting PPIs, it is difficult to predict structure target protein-cyclic peptide complex or design cyclic sequence that binds protein using computational methods. Recently, AlphaFold with offset has enabled predicting peptides, thereby enabling de novo designs. We developed enable structural prediction proteins and complexes found AlphaFold2 can structures high accuracy. also applied binder hallucination protocol AfDesign, method AlphaFold, we could predicted local-distance difference test lower separated binding energy per unit interface area native MDM2/p53 structure. Furthermore, was 12 other protein-peptide one complex. Our approach shows possible putative sequences PPI.

Language: Английский

Citations

17

Targeting protein-protein interactions with low molecular weight and short peptide modulators: insights on disease pathways and starting points for drug discovery DOI
Daniela Trisciuzzi, Bruno O. Villoutreix, Lydia Siragusa

et al.

Expert Opinion on Drug Discovery, Journal Year: 2023, Volume and Issue: 18(7), P. 737 - 752

Published: May 29, 2023

Introduction Protein-protein interactions (PPIs) have been often considered undruggable targets although they are attractive for the discovery of new therapeutics. The spread artificial intelligence and machine learning complemented with experimental methods is likely to change perspectives protein-protein modulator research. Noteworthy, some novel low molecular weight (LMW) short peptide modulators PPIs already in clinical trials treatment relevant diseases.Areas covered This review focuses on main properties interfaces key concepts pertaining modulation PPIs. authors survey recently reported state-of-the-art dealing rational design PPI highlight role several computer-based approaches.Expert opinion Interfering specifically large protein still an open challenge. initial concerns about unfavorable physicochemical many these nowadays less acute molecules lying beyond rule 5, orally available successful trials. As cost biologics interfering very high, it would seem reasonable put more effort, both academia private sectors, actively developing compounds peptides perform this task.

Language: Английский

Citations

13

Natural polymers as potential P-glycoprotein inhibitors: Pre-ADMET profile and computational analysis as a proof of concept to fight multidrug resistance in cancer DOI Creative Commons
Kumaraswamy Gandla, Fahadul Islam, Mehrukh Zehravi

et al.

Heliyon, Journal Year: 2023, Volume and Issue: 9(9), P. e19454 - e19454

Published: Aug. 24, 2023

P-glycoprotein (P-gp) is known as the "multidrug resistance protein" because it contributes to tumor several different classes of anticancer drugs. The effectiveness such polymers in treating cancer and delivering drugs has been shown a wide range vitro vivo experiments. primary objective present study was investigate inhibitory effects naturally occurring on P-gp efflux, that inhibition can impede elimination medications. our identify possess potential inhibit P-gp, protein involved drug resistance, with aim enhancing formulations. ADMET profile all selected (Agarose, Alginate, Carrageenan, Cyclodextrin, Dextran, Hyaluronic acid, Polysialic acid) studied, binding affinities were investigated through computational approach using recently released crystal structure PDB ID: 7O9W. advanced also done help molecular dynamics simulation. overcome MDR resulting from activity by when used docking scores native ligand, Agarose, Chitosan, acid found be −10.7, −8.5, −6.6, −8.7, −8.6, −24.5, −6.7, −8.3, −7.9, respectively. It observed that, Cyclodextrin multiple properties delivery science here demonstrated excellent affinity. We propose efflux-related may prevented use Carregeenan, and/or administration

Language: Английский

Citations

13

Fighting Antibiotic Resistance: New Pyrimidine-Clubbed Benzimidazole Derivatives as Potential DHFR Inhibitors DOI Creative Commons
M. Akiful Haque, Akash Marathakam, Ritesh Rana

et al.

Molecules, Journal Year: 2023, Volume and Issue: 28(2), P. 501 - 501

Published: Jan. 4, 2023

The present work describes the design and development of seventeen pyrimidine-clubbed benzimidazole derivatives as potential dihydrofolate reductase (DHFR) inhibitors. These compounds were filtered by using ADMET, drug-likeness characteristics calculations, molecular docking experiments. Compounds 27, 29, 30, 33, 37, 38, 41 chosen for synthesis based on results in silico screening. Each synthesized was tested its vitro antibacterial antifungal activities a variety strains. All showed properties against Gram-positive bacteria (Staphylococcus aureus Staphylococcus pyogenes) well Gram-negative (Escherichia coli Pseudomonas aeruginosa). Most either had higher potency than chloramphenicol or an equivalent to ciprofloxacin. 29 33 effective all bacterial fungal Finally, 1,2,3,4-tetrahydropyrimidine-2-thiol with 6-chloro-2-(chloromethyl)-1H-benzo[d]imidazole moiety are potent enough be considered promising lead discovery agent.

Language: Английский

Citations

10

Solubility-Aware Protein Binding Peptide Design Using AlphaFold DOI Creative Commons
Takatsugu Kosugi, Masahito Ohue

Biomedicines, Journal Year: 2022, Volume and Issue: 10(7), P. 1626 - 1626

Published: July 7, 2022

New protein-protein interactions (PPIs) are identified, but PPIs have different physicochemical properties compared with conventional targets, making it difficult to use small molecules. Peptides offer a new modality target PPIs, designing appropriate peptide sequences by computation is challenging. Recently, AlphaFold and RoseTTAFold made possible predict protein structures from amino acid ultra-high accuracy, enabling de novo design. We designed peptides likely PPI as the using "binder hallucination" protocol of AfDesign, design method AlphaFold. However, solubility tended be low. Therefore, we loss function indices for acids developed solubility-aware AfDesign binder hallucination protocol. The in increased weight function; moreover, they captured characteristics indices. Moreover, higher affinity than random or single residue substitution when evaluated docking binding affinity. Our approach shows that can bind interface while controlling solubility.

Language: Английский

Citations

14

AAontology: An Ontology of Amino Acid Scales for Interpretable Machine Learning DOI Creative Commons
Stephan Breimann, Frits Kamp, Harald Steiner

et al.

Journal of Molecular Biology, Journal Year: 2024, Volume and Issue: 436(19), P. 168717 - 168717

Published: July 24, 2024

Amino acid scales are crucial for protein prediction tasks, many of them being curated in the AAindex database. Despite various clustering attempts to organize and better understand their relationships, these approaches lack fine-grained classification necessary satisfactory interpretability problems. To address this issue, we developed AAontology—a two-level 586 amino (mainly from AAindex) together with an in-depth analysis relations—using bag-of-word-based classification, clustering, manual refinement over multiple iterations. AAontology organizes physicochemical into 8 categories 67 subcategories, enhancing scale-based machine learning methods bioinformatics. Thereby it enables researchers gain a deeper biological insight. We anticipate that will be building block link properties function dysfunctions as well aid informed decision-making mutation or drug design.

Language: Английский

Citations

3