Revolutionizing adjuvant development: harnessing AI for next-generation cancer vaccines

Frontiers in Immunology, Journal Year: 2024, Volume and Issue: 15

Published: Aug. 14, 2024

With the COVID-19 pandemic, importance of vaccines has been widely recognized and led to increased research development efforts. Vaccines also play a crucial role in cancer treatment by activating immune system target destroy cells. However, enhancing efficacy remains challenge. Adjuvants, which enhance response antigens improve vaccine effectiveness, have faced limitations recent years, resulting few novel adjuvants being identified. The advancement artificial intelligence (AI) technology drug provided foundation for adjuvant screening application, leading diversification adjuvants. This article reviews significant tumor basic clinical explores use AI screen from databases. findings this review offer valuable insights new next-generation vaccines.

Language: Английский

---

DiffBindFR: an SE(3) equivariant network for flexible protein–ligand docking

Chemical Science, Journal Year: 2024, Volume and Issue: 15(21), P. 7926 - 7942

Published: Jan. 1, 2024

DiffBindFR, a diffusion model based flexible full-atom protein–ligand docking tool, demonstrates its superior and side-chain refinement accuracy with reliable physical plausibility.

Language: Английский

---

A High-Quality Data Set of Protein–Ligand Binding Interactions Via Comparative Complex Structure Modeling

Journal of Chemical Information and Modeling, Journal Year: 2024, Volume and Issue: 64(7), P. 2454 - 2466

Published: Jan. 5, 2024

High-quality protein–ligand complex structures provide the basis for understanding nature of noncovalent binding interactions at atomic level and enable structure-based drug design. However, experimentally determined are scarce compared with vast chemical space. In this study, we addressed issue by constructing BindingNet data set via comparative structure modeling, which contains 69,816 modeled high-quality experimental affinity data. provides valuable insights into investigating interactions, allowing visual inspection interpretation structural analogues' structure–activity relationships. It can also be used evaluating machine-learning-based scoring functions. Our results indicate that machine learning models trained on could reduce bias caused buried solvent-accessible surface area, as previously found PDBbind set. We discussed strategies to improve its potential utilization benchmarking molecular docking methods ligand free energy calculation approaches. The complements in a sufficient unbiased is freely available http://bindingnet.huanglab.org.cn.

Language: Английский

---

Design and synthesis of a potential selective JAK-3 inhibitor for the treatment of rheumatoid arthritis using predictive QSAR models

Informatics in Medicine Unlocked, Journal Year: 2024, Volume and Issue: 45, P. 101464 - 101464

Published: Jan. 1, 2024

Language: Английский

---

A computational discovery of hexokinase 2 inhibitors from Newbouldia laevis for Hepatocellular carcinoma (HCC) treatment

South African Journal of Botany, Journal Year: 2024, Volume and Issue: 169, P. 12 - 26

Published: April 13, 2024

Newbouldia laevis, also known as the African Border Tree or Fever Tree, is a deciduous tree native to West Africa. The plant valued for its medicinal properties and used in traditional medicine antimicrobial anti-inflammatory effects. N. storage tank of phytochemicals with huge health benefits performances globally treatment management numerous disease conditions. Limited research exists on usage laevis hepatocellular carcinoma (HCC) treatment. This study aims explore inhibitory activities from against hexokinase 2 protein, target hepatocarcinoma presents unique silico approach that includes ligand binding site prediction, molecular docking, dynamics simulation, Molecular Mechanics Poisson–Boltzmann Surface Area (MM/PBSA) methods. A total 35 available 3D structures were identified through literature mining. review highlighted significance protein inhibitor (HCC). docking experiment all revealed had potential protein. Moreover, chrysarobin, apigenin ursolic acid best inhibitors lowest energy −8.9 kcal/mol, −8.7 −8.5 respectively. was validated by comparing affinities reference drug Cabozantinib-S-malate (−8.3 kcal/mol). Further, studies complexes scores described detail here. results 100 ns modeling (RMSD, RMSF, Rg SASA) show extraordinary stability during establishment acid, well favorable energy, which determined theoretically means MM/PBSA method, thereby increase probability their acting promising likely inhibitors. Therefore, predicted could be inhibitor/antagonist enzyme.

Language: Английский

---

Augmented BindingNet dataset for enhanced ligand binding pose predictions using deep learning

npj Drug Discovery., Journal Year: 2025, Volume and Issue: 2(1)

Published: Jan. 8, 2025

High-quality data on protein-ligand complex structures and binding affinities are crucial for structure-based drug design. Existing datasets often lack diversity quantity, limiting the comprehensive understanding of interactions. Here, we present BindingNet v2, an expanded dataset comprising 689,796 modeled complexes across 1794 protein targets. Constructed using enhanced template-based modeling workflow from v1, it incorporates pharmacophore molecular shape similarities. v2's effectiveness in pose generation was evaluated, showing improved generalization ability Uni-Mol model novel ligands. The success rate PoseBusters increased 38.55% with PDBbind alone to 64.25% augmenting v2. Coupled physics-based refinement, rose 74.07%, passing validity checks. These results highlight value larger, diverse enhancing accuracy reliability deep learning models prediction.

Language: Английский

---

A database for large-scale docking and experimental results

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 27, 2025

The rapid expansion of readily accessible compounds over the past six years has transformed molecular docking, improving hit rates and affinities. While many millions molecules may score well in a docking campaign, results are rarely fully shared, hindering benchmarking machine learning chemical space exploration methods that seek to explore expanding spaces. To address this gap, we develop website providing access recent large library campaigns, including poses, scores, vitro for campaigns against 11 targets, with 6.3 billion docked 3729 experimentally tested. In simple proof-of-concept study speaks new library's utility, use database train models predict scores find top 0.01% scoring while evaluating only 1% library. Even these studies, some interesting trends emerge: unsurprisingly, as on larger sets, they perform better; less expected, could achieve high correlations yet still fail enrich docking-discovered ligands, or even docking-ranked molecules. It will be see how more sophisticated than studies undertaken here; is openly available at lsd.docking.org.

Language: Английский

---

Generative adversarial neural network with graph embeddings for de novo designing small-molecule inhibitors against Mycobacterium tuberculosis KasA enzyme

Doklady of the National Academy of Sciences of Belarus, Journal Year: 2025, Volume and Issue: 69(1), P. 13 - 22

Published: Feb. 26, 2025

A generative semi-supervised adversarial neural network trained on graph embeddings was developed for de novo design of potential inhibitors against beta-ketoacyl-[acyl-carrier protein] synthase I (KasA), an enzyme critically important biosynthesis mycolic acids the Mycobacterium tuberculosis cell wall. The designed model and tested a set compounds from virtual library small molecules containing structural elements capable selective interactions with therapeutic target. Using network, 3,637 were designed, followed by assessment their inhibitory activity KasA protein using molecular docking methods. Based analysis obtained data, six exhibiting high affinity to malonyl-binding site selected. identified are assumed form promising basic structures further theoretical experimental studies development new effective drug-resistant tuberculosis.

Language: Английский

---

A Database for Large-Scale Docking and Experimental Results

Journal of Chemical Information and Modeling, Journal Year: 2025, Volume and Issue: unknown

Published: April 24, 2025

The rapid expansion of readily accessible compounds over the past six years has transformed molecular docking, improving hit rates and affinities. While many millions molecules may score well in a docking campaign, results are rarely fully shared, hindering benchmarking machine learning chemical space exploration methods that seek to explore expanding spaces. To address this gap, we develop website providing access recent large library campaigns, including poses, scores, vitro for campaigns against 11 targets, with 6.3 billion docked 3729 experimentally tested. In simple proof-of-concept study speaks new library's utility, use database train models predict scores find top 0.01% scoring while evaluating only 1% library. Even these studies, some interesting trends emerge: unsurprisingly, as on larger sets, they perform better; less expectedly, could achieve high correlations yet still fail enrich docking-discovered ligands, or even docking-ranked molecules. It will be see how more sophisticated than studies undertaken here; is openly available at lsd.docking.org.

Language: Английский

---

Identification of Novel HPK1 Hit Inhibitors: From In Silico Design to In Vitro Validation

International Journal of Molecular Sciences, Journal Year: 2025, Volume and Issue: 26(9), P. 4366 - 4366

Published: May 4, 2025

Hematopoietic progenitor kinase 1 (HPK1), a negative regulator of T-cells, B-cells, and dendritic cells, has gained attention in antitumor immunotherapy research over the past decade. No HPK1 inhibitor yet reached clinical approval, largely due to selectivity drug-like limitations. Leveraging available structural insights into HPK1, we conducted rational hit identification using structure-based virtual screening 600,000 molecules from ASINEX OTAVA databases. A series molecular docking studies, vitro assays, dynamics simulations were identify viable hits. This approach resulted two promising novel scaffolds, 4H-Pyrido[1,2-a] thieno[2,3-d] pyrimidin-4-one (ISR-05) quinolin-2(1H)-one (ISR-03), neither which previously been reported as an inhibitor. ISR-05 ISR-03 exhibited IC50 values 24.2 ± 5.07 43.9 0.134 µM, respectively, inhibition assays. These hits constitute tractable starting points for future hit-to-lead optimization aimed at developing more effective inhibitors cancer therapy.

Language: Английский

---