Structure-based Virtual Screening Research Articles

In silico identification of a novel Cdc2‐like kinase 2 (CLK2) inhibitor in triple negative breast cancer

AbstractDysregulation of RNA splicing processes is intricately linked to tumorigenesis in various cancers, especially breast cancer. Cdc2‐like kinase 2 (CLK2), an oncogenic RNA‐splicing kinase pivotal in breast cancer, plays a significant role, particularly in the context of triple‐negative breast cancer (TNBC), a subtype marked by substantial medical challenges due to its low survival rates. In this study, we employed a structure‐based virtual screening (SBVS) method to identify potential CLK2 inhibitors with novel chemical structures for treating TNBC. Compound 670551 emerged as a novel CLK2 inhibitor with a 50% inhibitory concentration (IC50) value of 619.7 nM. Importantly, Compound 670551 exhibited high selectivity for CLK2 over other protein kinases. Functionally, this compound significantly reduced the survival and proliferation of TNBC cells. Results from a cell‐based assay demonstrated that this inhibitor led to a decrease in RNA splicing proteins, such as SRSF4 and SRSF6, resulting in cell apoptosis. In summary, we identified a novel CLK2 inhibitor as a promising potential treatment for TNBC therapy.

Structure-Based Virtual Screening Identifying Novel FOXM1 Inhibitors as the Lead Compounds for Glioblastoma

Background: Glioblastoma multiforme (GBM) is a highly heterogeneous brain tumor with limited treatment options and a poor prognosis. Cancer stem cells (CSCs) have emerged as a critical factor in GBM resistance and management, contributing to tumor growth, heterogeneity, and immunosuppression. The transcription factor FOXM1 has been identified as a key player in the progression, spread, and therapy resistance of various cancers, including GBM. Objective: In this research, the objective was to perform structure-based in silico screening with the aim of identifying natural compounds proficient in targeting the DNA-binding domain (DBD) of the FOXM1 protein. Methods: In this study, in silico tools were employed for screening a hundred naturally occurring compounds capable of targeting the FOXM1 protein. Through molecular docking analysis and pharmacokinetic profiling, five compounds were found to be promising candidates for extensive interaction with the FOXM1 protein. Further, these compounds were validated for the stability of the FOXM1-natural compound complex using molecular dynamics (MD) simulations. Results: Four compounds, such as Withaferin A, Bryophyllin A, Silybin B, Sanguinarine and Troglitazone (control compound), emerged as promising candidates with substantial interactions with FOXM1, suggesting their potential as a protein inhibitor based on molecular docking investigations. After MD simulation analysis, the FOXM1- Bryophyllin A complex was found to maintain the highest stability, and the other three ligands had moderate but comparable binding affinities over a period of 100 ns. Conclusion: This study provides valuable insights into four promising FOXM1 inhibitors that have the ability to induce senescence in GBM stem cells. These findings contribute to the development of structure-based designing strategies for FOXM1 inhibitors and innovative therapeutic approaches for the treatment of Glioblastoma.

EGC enhances tumor antigen presentation and CD8+ T cell-mediated antitumor immunity via targeting oncoprotein SND1

The Staphylococcal nuclease and Tudor domain containing 1 (SND1) has been identified as an oncoprotein. Our previous study demonstrated that SND1 impedes the major histocompatibility complex class I (MHC-I) assembly by hijacking the nascent heavy chain of MHC-I to endoplasmic reticulum-associated degradation. Herein, we aimed to identify inhibitors to block SND1-MHC-I binding, to facilitate the MHC-I presentation and tumor immunotherapy. Our findings validated the importance of the K490-containing sites in SND1-MHC-I complex. Through structure-based virtual screening and docking analysis, (−)-Epigallocatechin (EGC) exhibited the highest docking score to prevent the binding of MHC-I to SND1 by altering the spatial conformation of SND1. Additionally, EGC treatment resulted in increased expression levels of membrane-presented MHC-I in tumor cells. The C57BL/6J murine orthotopic melanoma model validated that EGC increases infiltration and activity of CD8+ T cells in both the tumor and spleen. Furthermore, the combination of EGC with programmed death-1 (PD-1) antibody demonstrated a superior antitumor effect. In summary, we identified EGC as a novel inhibitor of SND1-MHC-I interaction, prompting MHC-I presentation to improve CD8+ T cell response within the tumor microenvironment. This discovery presents a promising immunotherapeutic candidate for tumors.

α-Glucosidase inhibitory potential of Oroxylum indicum using molecular docking, molecular dynamics, and in vitro evaluation

BackgroundAccording to the International Diabetes Federation, there will be 578 million individuals worldwide with diabetes by 2030 and 700 million by 2045. One of the promising drug targets to fight diabetes is α-glucosidase (AG), and its inhibitors may be used to manage diabetes by reducing the breakdown of complex carbohydrates into simple sugars. The study aims to identify and validate potential AG inhibitors in natural sources to combat diabetes. MethodsComputational techniques such as structure-based virtual screening and molecular dyncamic simulation were employed to predict potential AG inhibitors from compounds of Oroxylum indicum. Finally, in silico results were validated by in vitro analysis using n-butanol fraction of crude methanol extracts. ResultsThe XP glide scores of top seven hits OI_13, OI_66, OI_16, OI_44, OI_43, OI_20, OI_78 and acarbose were –14.261, –13.475, –13.074, –13.045, –12.978, –12.659, –12.354 and –12.296 kcal/mol, respectively. These hits demonstrated excellent binding affinity towards AG, surpassing the known AG inhibitor acarbose. The MM-GBSA dG binding energies of OI_13, OI_66, and acarbose were −69.093, −62.950, and −53.055 kcal/mol, respectively. Most of the top hits were glycosides, indicating that active compounds lie in the n-butanol fraction of the extract. The IC50 value for AG inhibition by n-butanol fraction was 248.1 μg/ml, and for that of pure acarbose it was 89.16 μg/ml. The predicted oral absorption rate in humans for the top seven hits was low like acarbose, which favors the use of these compounds as anti-diabetes in the small intestine. ConclusionIn summary, the study provides promising insights into the use of natural compounds derived from O. indicum as potential AG inhibitors to manage diabetes. However, further research, including clinical trials and pharmacological studies, would be necessary to validate their efficacy and safety before clinical use.

Antibiotic discovery with artificial intelligence for the treatment of Acinetobacter baumannii infections.

Acinetobacter baumannii presents a severe global health threat, with alarming levels of antimicrobial resistance rates resulting in significant morbidity and mortality in the USA, ranging from 26% to 68%, as reported by the Centers for Disease Control and Prevention (CDC). To address this threat, novel strategies beyond traditional antibiotics are imperative. Computational approaches, such as QSAR models leverage molecular structures to predict biological effects, expediting drug discovery. We identified OmpW as a potential therapeutic target in A. baumannii and screened 11,648 natural compounds. We employed QSAR models from a ChEMBL bioactivity data set and conducted structure-based virtual screening against OmpW. Demethoxycurcumin, a lead compound, exhibited promising antibacterial activity against A. baumannii, including multidrug-resistant strains. Additionally, demethoxycurcumin demonstrated anti-virulence properties by reducing A. baumannii interaction with host cells. The findings highlight the potential of artificial intelligence in discovering curcuminoids as effective antimicrobial agents against A. baumannii infections, offering a promising strategy to address antibiotic resistance.

In silico Identification of HDAC Inhibitors for Multiple Myeloma: A Structure-based Virtual Screening, Drug Likeness, ADMET Profiling, Molecular Docking, and Molecular Dynamics Simulation Study

Background: Multiple myeloma (MM) is a hematological malignancy of plasma cells that produce a monoclonal immunoglobulin protein. Despite significant advances in the treatment of MM, currently available therapies are associated with toxicity and resistance. As a result, there is an increasing demand for novel, effective therapeutics. Inhibition of histone deacetylases (HDACs) is emerging as a potential method for treating cancer. HDAC6 is one of 18 different HDAC isoforms that regulate tubulin lysine 40 and function in the microtubule network. HDAC6 participates in tumorigenesis and metastasis through protein ubiquitination, tubulin, and Hsp90. Several studies have found that inhibiting HDAC6 causes AKT and ERK dephosphorylation, which leads to decreased cell proliferation and promotes cancer cell death via the PI3K/AKT and MAPK/ERK signaling pathways. Objective: The objective of this study is to target HDAC6 and identify potent inhibitors for the treatment of multiple myeloma by employing computer-aided drug design. Materials and Methods: A total of 199,611,439 molecules from five different chemical databases, such as CHEMBL25, ChemSpace, Mcule, MolPort, and ZINC, have been screened against HDAC6 by structure- based virtual screening, followed by filtering for various drug-likeness, ADME, toxicity, consensus molecular docking, and 100 ns MD simulation. Results: Our research work resulted in three molecules that have shown strong binding affinity (CHEMBL2425964 -9.99 kcal/mol, CHEMBL2425966 -9.89 kcal/mol, and CSC067477144 -9.86 kcal/mol) at the active site HDAC6, along with effective ADME properties, low toxicity, and high stability. Inhibiting HDAC6 with these identified molecules will induce AKT and ERK dephosphorylation linked to reduced cell proliferation and promote cancer cell death. Conclusion: CHEMBL2425964, CHEMBL2425966, and CSC067477144 could be effective against multiple myeloma.

Ligand- and Structure-Based Virtual Screening Identifies New Inhibitors of the Interaction of the SARS-CoV-2 Spike Protein with the ACE2 Host Receptor

The Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) is a fast-spreading viral pathogen and poses a serious threat to human health. New SARS-CoV-2 variants have been arising worldwide; therefore, is necessary to explore more therapeutic options. The interaction of the viral spike (S) protein with the angiotensin-converting enzyme 2 (ACE2) host receptor is an attractive drug target to prevent the infection via the inhibition of virus cell entry. In this study, Ligand- and Structure-Based Virtual Screening (LBVS and SBVS) was performed to propose potential inhibitors capable of blocking the S receptor-binding domain (RBD) and ACE2 interaction. The best five lead compounds were confirmed as inhibitors through ELISA-based enzyme assays. The docking studies and molecular dynamic (MD) simulations of the selected compounds maintained the molecular interaction and stability (RMSD fluctuations less than 5 Å) with key residues of the S protein. The compounds DRI-1, DRI-2, DRI-3, DRI-4, and DRI-5 efficiently block the interaction between the SARS-CoV-2 spike protein and receptor ACE2 (from 69.90 to 99.65% of inhibition) at 50 µM. The most potent inhibitors were DRI-2 (IC50 = 8.8 µM) and DRI-3 (IC50 = 2.1 µM) and have an acceptable profile of cytotoxicity (CC50 > 90 µM). Therefore, these compounds could be good candidates for further SARS-CoV-2 preclinical experiments.

Structure-Based Virtual Screening of Antiviral Compounds Targeting the Norovirus RdRp Protein

Background: Human noroviruses (NV) are the primary etiological organisms causing acute gastroenteritis around the world, causing severe morbidity and imposing a significant economic burden. The RNA-dependent RNA polymerase (RdRp) is essential for viral replication and could be a promising target for anti-NV therapeutics. Despite the discovery of a few NV RdRp inhibitors, the majority of these pharmaceuticals have demonstrated limited efficacy in inhibiting viral replication in cellular models.Methods: In this study, computational screening of antiviral compounds was conducted targeting the NV RdRp protein. The assessment was based on binding poses and the key residues of RdRp involved in interactions with compounds.Results: The compounds namely, Ribavirin, BMS806, Dihydromyricetin, R7935788, and LY2784544 were found to bind the RdRp protein with high affinity. Notably, these compounds displayed significantly lower binding affinities compared to the positive control, PPNDS. In addition, these compounds exhibited many RdRp protein binding residues that were also present in the PPNDS.Conclusion: The results presented here suggest that these compounds have the potential to be used as inhibitors of NV RdRp in the development of antiviral medications. Nevertheless, due to the computational nature of this study, it is imperative to do experimental validation.Keywords: Noroviruses; RdRp; Virtual screening; Antiviral Compounds

Discovery of Novel Aldose Reductase Inhibitors via the Integration of Ligand-Based and Structure-Based Virtual Screening with Experimental Validation.

Aldose reductase plays a central role in diabetes mellitus (DM) associated complications by converting glucose to sorbitol, resulting in a harmful increase of reactive oxygen species (ROS) in various tissues, such as the heart, vasculature, neurons, eyes, and kidneys. We employed a comprehensive approach, integrating both ligand- and structure-based virtual screening followed by experimental validation. Initially, candidate compounds were extracted from extensive drug and chemical libraries using the DeepChem's GraphConvMol algorithm, leveraging its capacity for robust molecular feature representation. Subsequent refinement employed molecular docking and molecular dynamics (MD) simulations, which are crucial for understanding compound-receptor interactions and dynamic behavior in a simulated physiological environment. Finally, the candidate compounds were subjected to experimental validation of their biological activity using an aldose reductase inhibitor screening kit. The comprehensive approach led to the identification of a promising compound, demonstrating significant potential as an aldose reductase inhibitor. This comprehensive approach not only yields a potential therapeutic intervention for DM-related complications but also establishes an integrated protocol for drug development, setting a new benchmark in the field.

Structure-based virtual screening of mangiferin derivatives with antidiabetic action: a molecular docking and dynamics study and MPO-based drug-likeness approach.

The online version contains supplementary material available at 10.1007/s13205-024-03978-9.

Dexrazoxane inhibits the growth of esophageal squamous cell carcinoma by attenuating SDCBP/MDA-9/syntenin-mediated EGFR-PI3K-Akt pathway activation

Syndecan-binding protein (SDCBP) was reported to stimulate the advancement of esophageal squamous cell carcinoma (ESCC) and could potentially be a target for ESCC treatment. There is a growing corpus of research on the anti-tumor effects of iron chelators; however, very few studies have addressed the involvement of dexrazoxane in cancer. In this study, structure-based virtual screening was employed to select drugs targeting SDCBP from the Food and Drug Administration (FDA)-approved drug databases. The sepharose 4B beads pull-down assay revealed that dexrazoxane targeted SDCBP by interacting with its PDZ1 domain. Additionally, dexrazoxane inhibited ESCC cell proliferation and anchorage-independent colony formation via SDCBP. ESCC cell apoptosis and G2 phase arrest were induced as measured by the flow cytometry assay. Subsequent research revealed that dexrazoxane attenuated the binding ability between SDCBP and EGFR in an immunoprecipitation assay. Furthermore, dexrazoxane impaired EGFR membrane localization and inactivated the EGFR/PI3K/Akt pathway. In vivo, xenograft mouse experiments indicated that dexrazoxane suppressed ESCC tumor growth. These data indicate that dexrazoxane might be established as a potential anti-cancer agent in ESCC by targeting SDCBP.

Identification of naphthalimide-derivatives as novel PBD-targeted polo-like kinase 1 inhibitors with efficacy in drug-resistant lung cancer cells

Targeting polo-box domain (PBD) small molecule for polo-like kinase 1 (PLK1) inhibition is a viable alternative to target kinase domain (KD), which could avoid pan-selectivity and dose-limiting toxicity of ATP-competitive inhibitors. However, their efficacy in these settings is still low and inaccessible to clinical requirement. Herein, we utilized a structure-based high-throughput virtual screen to find novel chemical scaffold capable of inhibiting PLK1 via targeting PBD and identified an initial hit molecule compound 1a. Based on the lead compound 1a, a structural optimization approach was carried out and several series of derivatives with naphthalimide structural motif were synthesized. Compound 4Bb was identified as a new potent PLK1 inhibitor with a KD value of 0.29 μM. 4Bb could target PLK1 PBD to inhibit PLK1 activity and subsequently suppress the interaction of PLK1 with protein regulator of cytokinesis 1 (PRC1), finally leading to mitotic catastrophe in drug-resistant lung cancer cells. Furthermore, 4Bb could undergo nucleophilic substitution with the thiol group of glutathione (GSH) to disturb the redox homeostasis through exhausting GSH. By regulating cell cycle machinery and increasing cellular oxidative stress, 4Bb exhibited potent cytotoxicity to multiple cancer cells and drug-resistant cancer cells. Subcutaneous and oral administration of 4Bb could effectively inhibit the growth of drug-resistant tumors in vivo, doubling the survival time of tumor bearing mice without side effects in normal tissues. Thus, our study offers an orally-available, structurally-novel PLK1 inhibitor for drug-resistant lung cancer therapy.

CX-6258 hydrochloride hydrate: A potential non-nucleoside inhibitor targeting the RNA-dependent RNA polymerase of norovirus

Human norovirus (HuNoV), a primary cause of non-bacterial gastroenteritis, currently lacks approved treatment. RdRp is vital for virus replication, making it an attractive target for therapeutic intervention. By application of structure-based virtual screening procedure, we present CX-6258 hydrochloride hydrate as a potent RdRp non-nucleoside inhibitor, effectively inhibiting HuNoV RdRp activity with an IC50 of 3.61 μM. Importantly, this compound inhibits viral replication in cell culture, with an EC50 of 0.88 μM. In vitro binding assay validate that CX-6258 hydrochloride hydrate binds to RdRp through interaction with the “B-site” binding pocket. Interestingly, CX-6258-contacting residues such as R392, Q439, and Q414 are highly conserved among major norovirus GI and GII variants, suggesting that it may be a general inhibitor of norovirus RdRp. Given that CX-6258 hydrochloride hydrate is already utilized as an orally efficacious pan-Pim kinase inhibitor, it may serve as a potential lead compound in the effort to control HuNoV infections.

Design and Evaluation of NSAID Derivatives as AKR1C3 Inhibitors for Breast Cancer Treatment through Computer-Aided Drug Design and In Vitro Analysis.

Breast cancer is a major global health issue, causing high incidence and mortality rates as well as psychological stress for patients. Chemotherapy resistance is a common challenge, and the Aldo-keto reductase family one-member C3 enzyme is associated with resistance to anthracyclines like doxorubicin. Recent studies have identified celecoxib as a potential treatment for breast cancer. Virtual screening was conducted using a quantitative structure-activity relationship model to develop similar drugs; this involved backpropagation of artificial neural networks and structure-based virtual screening. The screening revealed that the C-6 molecule had a higher affinity for the enzyme (-11.4 kcal/mol), a lower half-maximal inhibitory concentration value (1.7 µM), and a safer toxicological profile than celecoxib. The compound C-6 was synthesized with an 82% yield, and its biological activity was evaluated. The results showed that C-6 had a more substantial cytotoxic effect on MCF-7 cells (62%) compared to DOX (63%) and celecoxib (79.5%). Additionally, C-6 had a less harmful impact on healthy L929 cells than DOX and celecoxib. These findings suggest that C-6 has promising potential as a breast cancer treatment.

Discovery and in vitro Evaluation of Novel Serotonin 5-HT2A Receptor Ligands Identified Through Virtual Screening.

The 5-HT2A receptor is a molecular target of high pharmacological importance. Ligands of this protein, particularly atypical antipsychotics, are useful in the treatment of numerous mental disorders, including schizophrenia and major depressive disorder. Structure-based virtual screening using a 5-HT2A receptor complex was performed to identify novel ligands for the 5-HT2A receptor, serving as potential antidepressants. From the Enamine screening library, containing over 4 million compounds, 48 molecules were selected for subsequent experimental validation. These compounds were tested against the 5-HT2A receptor in radioligand binding assays. From the tested batch, six molecules were identified as ligands of the main molecular target and were forwarded to a more detailed in vitro profiling. This included radioligand binding assays at 5-HT1A, 5-HT7, and D2 receptors and functional studies at 5-HT2A receptors. These compounds were confirmed to show a binding affinity for at least one of the targets tested in vitro. Thus, the success rate for the inactive template-based screening reached 17%, while it was 9% for the active template-based screening. Similarity and fragment analysis indicated the structural novelty of the identified compounds. Pharmacokinetics for these molecules was determined using in silico approaches.

Discovery of a novel SHP2 allosteric inhibitor using virtual screening, FMO calculation, and molecular dynamic simulation.

SHP2 is a non-receptor protein tyrosine phosphatase to remove tyrosine phosphorylation. Functionally, SHP2 is an essential bridge to connect numerous oncogenic cell-signaling cascades including RAS-ERK, PI3K-AKT, JAK-STAT, and PD-1/PD-L1 pathways. This study aims to discover novel and potent SHP2 inhibitors using a hierarchical structure-based virtual screening strategy that combines molecular docking and the fragment molecular orbital method (FMO) for calculating binding affinity (referred to as the Dock-FMO protocol). For the SHP2 target, the FMO method prediction has a high correlation between the binding affinity of the protein-ligand interaction and experimental values (R2 = 0.55), demonstrating a significant advantage over the MM/PBSA (R2 = 0.02) and MM/GBSA (R2 = 0.15) methods. Therefore, we employed Dock-FMO virtual screening of ChemDiv database of ∼2,990,000 compounds to identify a novel SHP2 allosteric inhibitor bearing hydroxyimino acetamide scaffold. Experimental validation demonstrated that the new compound (E)-2-(hydroxyimino)-2-phenyl-N-(piperidin-4-ylmethyl)acetamide (7188-0011) effectively inhibited SHP2 in a dose-dependent manner. Molecular dynamics (MD) simulation analysis revealed the binding stability of compound 7188-0011 and the SHP2 protein, along with the key interacting residues in the allosteric binding site. Overall, our work has identified a novel and promising allosteric inhibitor that targets SHP2, providing a new starting point for further optimization to develop more potent inhibitors. All the molecular docking studies were employed to identify potential leads with Maestro v10.1. The protein-ligand binding affinities of potential leads were further predicted by FMO calculations at MP2/6-31G* level using GAMESS v2020 system. MD simulations were carried out with AmberTools18 by applying the FF14SB force field. MD trajectories were analyzed using VMD v1.9.3. MM/GB(PB)SA binding free energy analysis was carried out with the mmpbsa.py tool of AmberTools18. The docking and MD simulation results were visualized through PyMOL v2.5.0.

Virtual Screening of a Chemically Diverse "Superscaffold" Library Enables Ligand Discovery for a Key GPCR Target.

The advent of ultra-large libraries of drug-like compounds has significantly broadened the possibilities in structure-based virtual screening, accelerating the discovery and optimization of high-quality lead chemotypes for diverse clinical targets. Compared to traditional high-throughput screening, which is constrained to libraries of approximately one million compounds, the ultra-large virtual screening approach offers substantial advantages in both cost and time efficiency. By expanding the chemical space with compounds synthesized from easily accessible and reproducible reactions and utilizing a large, diverse set of building blocks, we can enhance both the diversity and quality of the discovered lead chemotypes. In this study, we explore new chemical spaces using reactions of sulfur(VI) fluorides to create a combinatorial library consisting of several hundred million compounds. We screened this virtual library for cannabinoid type II receptor (CB2) antagonists using the high-resolution structure in conjunction with a rationally designed antagonist, AM10257. The top-predicted compounds were then synthesized and tested in vitro for CB2 binding and functional antagonism, achieving an experimentally validated hit rate of 55%. Our findings demonstrate the effectiveness of reliable reactions, such as sulfur fluoride exchange, in diversifying ultra-large chemical spaces and facilitate the discovery of new lead compounds for important biological targets.

Identification of small molecules affecting the interaction between human hemoglobin and Staphylococcus aureus IsdB hemophore

Human hemoglobin (Hb) is the preferred iron source of Staphylococcus aureus. This pathogenic bacterium exploits a sophisticated protein machinery called Iron-regulated surface determinant (Isd) system to bind Hb, extract and internalize heme, and finally degrade it to complete iron acquisition. IsdB, the surface exposed Hb receptor, is a proven virulence factor of S. aureus and the inhibition of its interaction with Hb can be pursued as a strategy to develop new classes of antimicrobials. To identify small molecules able to disrupt IsdB:Hb protein–protein interactions (PPIs), we carried out a structure-based virtual screening campaign and developed an ad hoc immunoassay to screen the retrieved set of commercially available compounds. Saturation-transfer difference (STD) NMR was applied to verify specific interactions of a sub-set of molecules, chosen based on their efficacy in reducing the amount of Hb bound to IsdB. Among molecules for which direct binding was verified, the best hit was submitted to ITC analysis to measure the binding affinity to Hb, which was found to be in the low micromolar range. The results demonstrate the viability of the proposed in silico/in vitro experimental pipeline to discover and test IsdB:Hb PPI inhibitors. The identified lead compound will be the starting point for future SAR and molecule optimization campaigns.

Comprehensive machine learning boosts structure-based virtual screening for PARP1 inhibitors

Poly ADP-ribose polymerase 1 (PARP1) is an attractive therapeutic target for cancer treatment. Machine-learning scoring functions constitute a promising approach to discovering novel PARP1 inhibitors. Cutting-edge PARP1-specific machine-learning scoring functions were investigated using semi-synthetic training data from docking activity-labelled molecules: known PARP1 inhibitors, hard-to-discriminate decoys property-matched to them with generative graph neural networks and confirmed inactives. We further made test sets harder by including only molecules dissimilar to those in the training set. Comprehensive analysis of these datasets using five supervised learning algorithms, and protein–ligand fingerprints extracted from docking poses and ligand only features revealed one highly predictive scoring function. This is the PARP1-specific support vector machine-based regressor, when employing PLEC fingerprints, which achieved a high Normalized Enrichment Factor at the top 1% on the hardest test set (NEF1% = 0.588, median of 10 repetitions), and was more predictive than any other investigated scoring function, especially the classical scoring function employed as baseline.

Advancing Ligand Docking through Deep Learning: Challenges and Prospects in Virtual Screening.

ConspectusMolecular docking, also termed ligand docking (LD), is a pivotal element of structure-based virtual screening (SBVS) used to predict the binding conformations and affinities of protein-ligand complexes. Traditional LD methodologies rely on a search and scoring framework, utilizing heuristic algorithms to explore binding conformations and scoring functions to evaluate binding strengths. However, to meet the efficiency demands of SBVS, these algorithms and functions are often simplified, prioritizing speed over accuracy.The emergence of deep learning (DL) has exerted a profound impact on diverse fields, ranging from natural language processing to computer vision and drug discovery. DeepMind's AlphaFold2 has impressively exhibited its ability to accurately predict protein structures solely from amino acid sequences, highlighting the remarkable potential of DL in conformation prediction. This groundbreaking advancement circumvents the traditional search-scoring frameworks in LD, enhancing both accuracy and processing speed and thereby catalyzing a broader adoption of DL algorithms in binding pose prediction. Nevertheless, a consensus on certain aspects remains elusive.In this Account, we delineate the current status of employing DL to augment LD within the VS paradigm, highlighting our contributions to this domain. Furthermore, we discuss the challenges and future prospects, drawing insights from our scholarly investigations. Initially, we present an overview of VS and LD, followed by an introduction to DL paradigms, which deviate significantly from traditional search-scoring frameworks. Subsequently, we delve into the challenges associated with the development of DL-based LD (DLLD), encompassing evaluation metrics, application scenarios, and physical plausibility of the predicted conformations. In the evaluation of LD algorithms, it is essential to recognize the multifaceted nature of the metrics. While the accuracy of binding pose prediction, often measured by the success rate, is a pivotal aspect, the scoring/screening power and computational speed of these algorithms are equally important given the pivotal role of LD tools in VS. Regarding application scenarios, early methods focused on blind docking, where the binding site is unknown. However, recent studies suggest a shift toward identifying binding sites rather than solely predicting binding poses within these models. In contrast, LD with a known pocket in VS has been shown to be more practical. Physical plausibility poses another significant challenge. Although DLLD models often achieve higher success rates compared to traditional methods, they may generate poses with implausible local structures, such as incorrect bond angles or lengths, which are disadvantageous for postprocessing tasks like visualization. Finally, we discuss the future perspectives for DLLD, emphasizing the need to improve generalization ability, strike a balance between speed and accuracy, account for protein conformation flexibility, and enhance physical plausibility. Additionally, we delve into the comparison between generative and regression algorithms in this context, exploring their respective strengths and potential.