Discovery of novel FGFR4 inhibitors through a build-up fragment strategy

Hepatocellular carcinoma (HCC) is a leading cause of cancer-related death. FGFR4 has been implicated in HCC progression, making it a promising therapeutic target. We introduce an approach for identifying novel FGFR4 inhibitors by sequentially adding fragments to a common warhead unit. This strategy resulted in the discovery of a potent inhibitor, 4c, with an IC50 of 33 nM and high selectivity among members of the FGFR family. Although further optimisation is required, our approach demonstrated the potential for discovering potent FGFR4 inhibitors for HCC treatment, and provides a useful method for obtaining hit compounds from small fragments.

Discovery of a potent inhibitor, D-132, targeting AsfvPolX, via protein-DNA complex-guided pharmacophore screening and in vitro molecular characterizations

The heightened transmissibility and capacity of African swine fever virus (ASFV) induce fatal diseases in domestic pigs and wild boars, posing significant economic repercussions and global threats. Despite extensive research efforts, the development of potent vaccines or treatments for ASFV remains a persistent challenge. Recently, inhibiting the AsfvPolX, a key DNA repair enzyme, emerges as a feasible strategy to disrupt viral replication and control ASFV infections. In this study, a comprehensive approach involving pharmacophore-based inhibitor screening, coupled with biochemical and biophysical analyses, were implemented to identify, characterize, and validate potential inhibitors targeting AsfvPolX. The constructed pharmacophore model, Phar-PolX-S, demonstrated efficacy in identifying a potent inhibitor, D-132 (IC50 = 2.8 ± 0.2 µM), disrupting the formation of the AsfvPolX-DNA complex. Notably, D-132 exhibited strong binding to AsfvPolX (KD = 6.9 ± 2.2 µM) through a slow-on-fast-off binding mechanism. Employing molecular modeling, it was elucidated that D-132 predominantly binds in-between the palm and finger domains of AsfvPolX, with crucial residues (R42, N48, Q98, E100, F102, and F116) identified as hotspots for structure-based inhibitor optimization. Distinctively characterized by a 1,2,5,6-tetrathiocane with modifications at the 3 and 8 positions involving ethanesulfonates, D-132 holds considerable promise as a lead compound for the development of innovative agents to combat ASFV infections.

Machine learning-guided design of potent darunavir analogs targeting HIV-1 proteases: A computational approach for antiretroviral drug discovery

In the pursuit of novel antiretroviral therapies for human immunodeficiency virus type-1 (HIV-1) proteases (PRs), recent improvements in drug discovery have embraced machine learning (ML) techniques to guide the design process. This study employs ensemble learning models to identify crucial substructures as significant features for drug development. Using molecular docking techniques, a collection of 160 darunavir (DRV) analogs was designed based on these key substructures and subsequently screened using molecular docking techniques. Chemical structures with high fitness scores were selected, combined, and one-dimensional (1D) screening based on beyond Lipinski's rule of five (bRo5) and ADME (absorption, distribution, metabolism, and excretion) prediction implemented in the Combined Analog generator Tool (CAT) program. A total of 473 screened analogs were subjected to docking analysis through convolutional neural networks scoring function against both the wild-type (WT) and 12 major mutated PRs. DRV analogs with negative changes in binding free energy (ΔΔ G bind ) compared to DRV could be categorized into four attractive groups based on their interactions with the majority of vital PRs. The analysis of interaction profiles revealed that potent designed analogs, targeting both WT and mutant PRs, exhibited interactions with common key amino acid residues. This observation further confirms that the ML model-guided approach effectively identified the substructures that play a crucial role in potent analogs. It is expected to function as a powerful computational tool, offering valuable guidance in the identification of chemical substructures for synthesis and subsequent experimental testing.

G-K BertDTA: A graph representation learning and semantic embedding-based framework for drug-target affinity prediction

Developing new drugs is costly, time-consuming, and risky. Drug-target affinity (DTA), indicating the binding capability between drugs and target proteins, is a crucial indicator for drug development. Accurately predicting interaction strength between new drug-target pairs by analyzing previous experiments aids in screening potential drug molecules, repurposing them, and developing safe and effective medicines. Existing computational models for DTA prediction rely on strings or single-graph neural networks, lacking consideration of protein structure and molecular semantic information, leading to limited accuracy. Our experiments demonstrate that string-based methods may overlook protein conformations, causing a high root mean square error (RMSE) of 3.584 in affinity due to a lack of spatial context. Single graph networks also underperform on topology features, with a 6% lower confidence interval (CI) for activity classification. Absent semantic information also limits generalization across diverse compounds, resulting in 18% increment in RMSE and 5% in misclassifications within quantifications study, restricting potential drug discovery. To address these limitations, we propose G-K BertDTA, a novel framework for accurate DTA prediction incorporating protein features, molecular semantic features, and molecular structural information. In this proposed model, we represent drugs as graphs, with a GIN employed to learn the molecular topological information. For the extraction of protein structural features, we utilize a DenseNet architecture. A knowledge-based BERT semantic model is incorporated to obtain rich pre-trained semantic embeddings, thereby enhancing the feature information. We extensively evaluated our proposed approach on the publicly available benchmark datasets (i.e., KIBA and Davis), and experimental results demonstrate the promising performance of our method, which consistently outperforms previous state-of-the-art approaches. Code is available at https://github.com/AmbitYuki/G-K-BertDTA.

Discovery of a new Bcl-2 inhibitor through synthesis, anticancer activity, docking and MD simulations

A database of 300 compounds was virtually screened and docked against Bcl-2 protein; the stability of the best-formed complex was evaluated through Molecular dynamics, the top ten compounds with the best in-silico complexation affinities were synthesized, and their In-vitro cytotoxic activity was examined. Thiazolidinone (4e) and isoxazoline (4a-d) were evaluated in-silico. For further evaluation and examination, we designed and synthesized from naturally occurring (R)-carvone and characterized it via spectroscopic analysis, as well as tested for their anticancer activities towards human cancer cell lines such as HT-1080 (fibrosarcome cancer), MCF-7 and MDA-MB-231 (breast cancer) and A-549 (lung cancer) by using MTT method with Doxorubicin as standard drug. Among them, compound 4d showed the most promising anticancer activity against HT-1080, A-549, MCF-7, and MDA-MB-231 cell lines with IC50 values of 15.59 ± 3.21 µM; 18.32 ± 2.73 µM; 17.28 ± 0.33 µM and 19.27 ± 2.73 µM respectively.Communicated by Ramaswamy H. Sarma.

The application of brain organoid for drug discovery in mitochondrial diseases

Mitochondrial diseases are difficult to treat due to the complexity and multifaceted nature of mitochondrial dysfunction. Brain organoids are three-dimensional (3D) structures derived from human pluripotent stem cells designed to mimic brain-like development and function. Brain organoids have received a lot of attention in recent years as powerful tools for modeling human diseases, brain development, and drug screening. Screening compounds for mitochondrial diseases using brain organoids could provide a more physiologically relevant platform for drug discovery. Brain organoids offer the possibility of personalized medicine because they can be derived from patient-specific cells, allowing testing of drugs tailored to specific genetic mutations. In this article, we highlight how brain organoids offer a promising avenue for screening compounds for mitochondrial diseases and address the challenges and limitations associated with their use. We hope this review will provide new insights into the further progress of brain organoids for mitochondrial screening studies.

C3-Symmetric ligands in drug design: An overview of the challenges and opportunities ahead

C3-symmetry is a type of star-shaped molecule consisting of a central core and three symmetrically attached chains. These molecules are used in drug discovery due to their unique three-fold rotational symmetry, which allows for specific binding interactions and improved molecular recognition. In this text, we provide an overview of synthetic approaches with C3-symmetry as a pharmaceutical tool: progress, challenges, and opportunities. C3-symmetric ligands offer both challenges and opportunities in drug design. Their unique symmetry can enhance binding interactions, but careful consideration of rigidity, synthetic complexity, and target compatibility is crucial. Further research and advancements in synthetic methods and modeling tools will likely drive their exploration in drug discovery, leading to the discovery of potent C3-symmetric ligands.

Newly uncovered Cryo-EM structures of mammalian NCXs set a new stage for resolving the underlying molecular mechanisms and drug discovery

The membrane-abundant NCX proteins mediate an electrogenic ion exchange (3Na+:1Ca2+) in the Ca2+-exit or Ca2+-entry mode. The structurally related isoform/splice variants of NCX are expressed in a tissue-specific manner to shape Ca2+ signalling/homeostasis in diverse cell types. The lack of mammalian NCX structure hampered the functional and regulatory resolution of tissue-specific NCX variants and their pharmacological targeting. Recently unveiled Cryo-EM structures of human cardiac NCX1.1[1] and kidney NCX1.3[2] provide new opportunities for resolving structure/functional divergences among NCX variants and their pharmacological targeting.

CXCR4/CXCL12 axis: "old" pathway as "novel" target for anti-inflammatory drug discovery

Inflammation is the body's defense response to exogenous or endogenous stimuli, involving complex regulatory mechanisms. Discovering anti-inflammatory drugs with both effectiveness and long-term use safety is still the direction of researchers' efforts. The inflammatory pathway was initially identified to be involved in tumor metastasis and HIV infection. However, research in recent years has proved that the CXC chemokine receptor type 4 (CXCR4)/CXC motif chemokine ligand 12 (CXCL12) axis plays a critical role in the upstream of the inflammatory pathway due to its chemotaxis to inflammatory cells. Blocking the chemotaxis of inflammatory cells by CXCL12 at the inflammatory site may block and alleviate the inflammatory response. Therefore, developing CXCR4 antagonists has become a novel strategy for anti-inflammatory therapy. This review aimed to systematically summarize and analyze the mechanisms of action of the CXCR4/CXCL12 axis in more than 20 inflammatory diseases, highlighting its crucial role in inflammation. Additionally, the anti-inflammatory activities of CXCR4 antagonists were discussed. The findings might help generate new perspectives for developing anti-inflammatory drugs targeting the CXCR4/CXCL12 axis.

What is the role of current mass spectrometry in pharmaceutical analysis?

The role of mass spectrometry (MS) has become more important in most application domains in recent years. Pharmaceutical analysis is specific due to its stringent regulation procedures, the need for good laboratory/manufacturing practices, and a large number of routine quality control analyses to be carried out. The role of MS is, therefore, very different throughout the whole drug development cycle. While it dominates within the drug discovery and development phase, in routine quality control, the role of MS is minor and indispensable only for selected applications. Moreover, its role is very different in the case of analysis of small molecule pharmaceuticals and biopharmaceuticals. Our review explains the role of current MS in the analysis of both small-molecule chemical drugs and biopharmaceuticals. Important features of MS-based technologies being implemented, method requirements, and related challenges are discussed. The differences in analytical procedures for small molecule pharmaceuticals and biopharmaceuticals are pointed out. While a single method or a small set of methods is usually sufficient for quality control in the case of small molecule pharmaceuticals and MS is often not indispensable, a large panel of methods including extensive use of MS must be used for quality control of biopharmaceuticals. Finally, expected development and future trends are outlined.

INTransformer: Data augmentation-based contrastive learning by injecting noise into transformer for molecular property prediction

Molecular property prediction plays an essential role in drug discovery for identifying the candidate molecules with target properties. Deep learning models usually require sufficient labeled data to train good prediction models. However, the size of labeled data is usually small for molecular property prediction, which brings great challenges to deep learning-based molecular property prediction methods. Furthermore, the global information of molecules is critical for predicting molecular properties. Therefore, we propose INTransformer for molecular property prediction, which is a data augmentation method via contrastive learning to alleviate the limitations of the labeled molecular data while enhancing the ability to capture global information. Specifically, INTransformer consists of two identical Transformer sub-encoders to extract the molecular representation from the original SMILES and noisy SMILES respectively, while achieving the goal of data augmentation. To reduce the influence of noise, we use contrastive learning to ensure the molecular encoding of noisy SMILES is consistent with that of the original input so that the molecular representation information can be better extracted by INTransformer. Experiments on various benchmark datasets show that INTransformer achieved competitive performance for molecular property prediction tasks compared with the baselines and state-of-the-art methods.

Discovery of macrocyclic CDK2/4/6 inhibitors with improved potency and DMPK properties through a highly efficient macrocyclic drug design platform

Cyclin-dependent kinases (CDKs) are critical cell cycle regulators that are often overexpressed in tumors, making them promising targets for anti-cancer therapies. Despite substantial advancements in optimizing the selectivity and drug-like properties of CDK inhibitors, safety of multi-target inhibitors remains a significant challenge. Macrocyclization is a promising drug discovery strategy to improve the pharmacological properties of existing compounds. Here we report the development of a macrocyclization platform that enabled the highly efficient discovery of a novel, macrocyclic CDK2/4/6 inhibitor from an acyclic precursor (NUV422). Using dihedral angle scan and structure-based, computer-aided drug design to select an optimal ring-closing site and linker length for the macrocycle, we identified compound 8 as a potent new CDK2/4/6 inhibitor with optimized cellular potency and safety profile compared to NUV422. Our platform leverages both experimentally-solved as well as generative chemistry-derived macrocyclic structures and can be deployed to streamline the design of macrocyclic new drugs from acyclic starting compounds, yielding macrocyclic compounds with enhanced potency and improved drug-like properties.

The preclinical discovery and clinical evaluation of tirzepatide for the treatment of type 2 diabetes

INTRODUCTION

Despite numerous antidiabetic medications available for the treatment of type 2 diabetes, a substantial percentage of patients fail to achieve optimal glycemic control. Furthermore, the escalating obesity pandemic underscores the urgent need for effective relevant pharmacotherapies. Tirzepatide, a novel dual GIP and GLP-1 receptor agonist, offers a promising therapeutic option.

AREAS COVERED

This review describes the discovery and clinical development of tirzepatide. Based on data from pivotal in vivo and in vitro studies, the authors present the pharmacodynamic profile of tirzepatide. Furthermore, they summarize data from the clinical trial programs that assessed the efficacy and safety of tirzepatide for the treatment of type 2 diabetes or obesity in a broad spectrum of patients, and discuss its therapeutic potential.

EXPERT OPINION

Tirzepatide effectively reduces glucose levels and body weight in patients with type 2 diabetes and/or obesity, with a generally safe profile. Based on data from phase 3 clinical trials, several agencies have approved its use for the treatment of type 2 diabetes and obesity. Clinicians should be aware of possible adverse events, mainly mild-to-moderate gastrointestinal side effects. Overall, tirzepatide represents a promising treatment option for the treatment of type 2 diabetes.

GraphormerDTI: A graph transformer-based approach for drug-target interaction prediction

The application of Artificial Intelligence (AI) to screen drug molecules with potential therapeutic effects has revolutionized the drug discovery process, with significantly lower economic cost and time consumption than the traditional drug discovery pipeline. With the great power of AI, it is possible to rapidly search the vast chemical space for potential drug-target interactions (DTIs) between candidate drug molecules and disease protein targets. However, only a small proportion of molecules have labelled DTIs, consequently limiting the performance of AI-based drug screening. To solve this problem, a machine learning-based approach with great ability to generalize DTI prediction across molecules is desirable. Many existing machine learning approaches for DTI identification failed to exploit the full information with respect to the topological structures of candidate molecules. To develop a better approach for DTI prediction, we propose GraphormerDTI, which employs the powerful Graph Transformer neural network to model molecular structures. GraphormerDTI embeds molecular graphs into vector-format representations through iterative Transformer-based message passing, which encodes molecules' structural characteristics by node centrality encoding, node spatial encoding and edge encoding. With a strong structural inductive bias, the proposed GraphormerDTI approach can effectively infer informative representations for out-of-sample molecules and as such, it is capable of predicting DTIs across molecules with an exceptional performance. GraphormerDTI integrates the Graph Transformer neural network with a 1-dimensional Convolutional Neural Network (1D-CNN) to extract the drugs' and target proteins' representations and leverages an attention mechanism to model the interactions between them. To examine GraphormerDTI's performance for DTI prediction, we conduct experiments on three benchmark datasets, where GraphormerDTI achieves a superior performance than five state-of-the-art baselines for out-of-molecule DTI prediction, including GNN-CPI, GNN-PT, DeepEmbedding-DTI, MolTrans and HyperAttentionDTI, and is on a par with the best baseline for transductive DTI prediction. The source codes and datasets are publicly accessible at https://github.com/mengmeng34/GraphormerDTI.

Discovery and Characterization of a Novel Cereblon-Recruiting PRC1 Bridged PROTAC Degrader

Bridged PROTAC is a novel protein complex degrader strategy that exploits the target protein's binding partner to degrade undruggable proteins by inducing proximity to an E3 ubiquitin ligase. In this study, we discovered for the first time that cereblon (CRBN) can be employed for the bridged PROTAC approach and report the first-in-class CRBN-recruiting and EED-binding polycomb repressive complex 1 (PRC1) degrader, compound 1 (MS181). We show that 1 induces preferential degradation of PRC1 components, BMI1 and RING1B, in an EED-, CRBN-, and ubiquitin-proteosome system (UPS)-dependent manner. Compound 1 also has superior antiproliferative activity in multiple metastatic cancer cell lines over EED-binding PRC2 degraders and can be efficacious in VHL-defective cancer cells. Altogether, compound 1 is a valuable chemical biology tool to study the role of PRC1 in cancer. Importantly, we show that CRBN can be utilized to develop bridged PROTACs, expanding the bridged PROTAC technology for degrading undruggable proteins.

Structure-Guided Discovery and Preclinical Assessment of Novel (Thiophen-3-yl)aminopyrimidine Derivatives as Potent ERK1/2 Inhibitors

The RAS-RAF-MEK-ERK signaling cascade is abnormally activated in various tumors, playing a crucial role in mediating tumor progression. As the key component at the terminal stage of this cascade, ERK1/2 emerges as a potential antitumor target and offers a promising therapeutic strategy for tumors harboring BRAF or RAS mutations. Here, we identified 36c with a (thiophen-3-yl)aminopyrimidine scaffold as a potent ERK1/2 inhibitor through structure-guided optimization for hit 18. In preclinical studies, 36c showed powerful ERK1/2 inhibitory activities (ERK1/2 IC50 = 0.11/0.08 nM) and potent antitumor efficacy both in vitro and in vivo against triple-negative breast cancer and colorectal cancer models harboring BRAF and RAS mutations. 36c could directly inhibit ERK1/2, significantly block the phosphorylation expression of their downstream substrates p90RSK and c-Myc, and induce cell apoptosis and incomplete autophagy-related cell death. Taken together, this work provides a promising ERK1/2 lead compound for multiple tumor-treatment drug discovery.

Discovery of a Novel CSF-1R Inhibitor with Highly Improved Pharmacokinetic Profiles and Superior Efficacy in Colorectal Cancer Immunotherapy

Blocking CSF-1/CSF-1R pathway has emerged as a promising strategy to remodel tumor immune microenvironment (TME) by reprogramming tumor-associated macrophages (TAMs). In this work, a novel CSF-1R inhibitor C19 with a highly improved pharmacokinetic profile and in vivo anticolorectal cancer (CRC) efficiency was successfully discovered. C19 could effectively reprogram M2-like TAMs to M1 phenotype and reshape the TME by inducing the recruitment of CD8+ T cells into tumors and reducing the infiltration of immunosuppressive Tregs/MDSCs. Deeper mechanistic studies revealed that C19 facilitated the infiltration of CD8+ T cells by enhancing the secretion of chemokine CXCL9, thus significantly potentiating the anti-CRC efficiency of PD-1 blockade. More importantly, C19 combined with PD-1 mAb could induce durable antitumor immune memory, effectively overcoming the recurrence of CRC. Taken together, our findings suggest that C19 is a promising therapeutic option for sensitizing CRC to anti-PD-1 therapy.

Discovery of Potent and Selective PI3Kδ Inhibitors for the Treatment of Acute Myeloid Leukemia

PI3Kδ is an essential target correlated to the occurrence and development of acute myeloid leukemia (AML). Herein, we investigated the pyrazolo[3,4-d]pyrimidine derivatives as potent and selective PI3Kδ inhibitors with high therapeutic efficacy toward AML. There were 44 compounds designed and prepared in a four-round optimization, and the biological evaluation showed that (S)-36 exhibited potent PI3Kδ inhibitory activity, high selectivity, and high antiproliferative activities against MV-4-11 and MOLM-13 cells, coupled with high oral bioavailability (F = 59.6%). In the MOLM-13 subcutaneous xenograft model, (S)-36 could significantly suppress the tumor progression with a TGI of 67.81% at an oral administration dosage of 10 mg/kg without exhibiting obvious toxicity. Mechanistically, (S)-36 could robustly inhibit the PI3K/AKT pathway for significant suppression of cell proliferation and remarkable induction of apoptosis both in vitro and in vivo. Thus, compound (S)-36 represents a promising PI3Kδ inhibitor for the treatment of acute myeloid leukemia with high efficacy.

Discovery of Novel Heterotricyclic Compounds as DNA-Dependent Protein Kinase (DNA-PK) Inhibitors with Enhanced Chemosensitivity, Oral Bioavailability, and the Ability to Potentiate Cancer Immunotherapy

In this work, a novel series of heterotricyclic DNA-PK inhibitors were rationally designed, synthesized, and assessed for their biological activity. In the DNA-PK biochemical assay, most compounds displayed potent enzymatic activity, with IC50 values between 0.11 and 71.5 nM. Among them, SK10 exhibited the most potent DNA-PK-inhibitory activity (IC50 = 0.11 nM). Studies of the mechanism of action indicated that SK10 could lower γH2A.X expression levels and demonstrate optimal synergistic antiproliferative activity against Jurkat cells (IC50 = 25 nM) when combined with doxorubicin. Importantly, in CT26 and B16-F10 tumor-bearing mouse models, the combination therapies of SK10 with chemotherapeutic drug doxorubicin, a PD-L1 antibody, and SWS1 (a potent PD-L1 small-molecule inhibitor) demonstrated superior synergistic anticancer and potential immunomodulatory effects. Furthermore, SK10 possessed favorable in vivo pharmacokinetic properties [e.g., oral bioavailability (F) = 31.8%]. Taken together, SK10 represents a novel heterotricyclic DNA-PK inhibitor with antitumor immune effects and favorable pharmacokinetics.

Discovery of 2-Amide-3-methylester Thiophenes that Target SARS-CoV-2 Mac1 and Repress Coronavirus Replication, Validating Mac1 as an Antiviral Target

The COVID-19 pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) virus has made it clear that further development of antiviral therapies will be needed. Here, we describe small-molecule inhibitors for SARS-CoV-2 Mac1, which counters ADP-ribosylation-mediated innate immune responses. Three high-throughput screening hits had the same 2-amide-3-methylester thiophene scaffold. We studied the compound binding mode using X-ray crystallography, allowing us to design analogues. Compound 27 (MDOLL-0229) had an IC50 of 2.1 μM and was selective for CoV Mac1 proteins after profiling for activity against a panel of viral and human proteins. The improved potency allowed testing of its effect on virus replication, and indeed, 27 inhibited replication of both murine hepatitis virus (MHV) prototypes CoV and SARS-CoV-2. Sequencing of a drug-resistant MHV identified mutations in Mac1, further demonstrating the specificity of 27. Compound 27 is the first Mac1-targeted small molecule demonstrated to inhibit coronavirus replication in a cell model.

Discovery of the First-in-Class G9a/GLP PROTAC Degrader

Aberrantly expressed lysine methyltransferases G9a and GLP, which catalyze mono- and dimethylation of histone H3 lysine 9 (H3K9), have been implicated in numerous cancers. Recent studies have uncovered both catalytic and noncatalytic oncogenic functions of G9a/GLP. As such, G9a/GLP catalytic inhibitors have displayed limited anticancer activity. Here, we report the discovery of the first-in-class G9a/GLP proteolysis targeting chimera (PROTAC) degrader 10 (MS8709), as a potential anticancer therapeutic. 10 induces G9a/GLP degradation in a concentration-, time-, and ubiquitin-proteasome system (UPS)-dependent manner. Futhermore, 10 does not alter the mRNA expression of G9a/GLP and is selective for G9a/GLP over other methyltransferases. Moreover, 10 displays superior cell growth inhibition to the parent G9a/GLP inhibitor UNC0642 in prostate, leukemia, and lung cancer cells and has suitable mouse pharmacokinetic properties for in vivo efficacy studies. Overall, 10 is a valuable chemical biology tool to further investigate the functions of G9a/GLP and a potential therapeutic for treating G9a/GLP-dependent cancers.

Discovery of YFJ-36: Design, Synthesis, and Antibacterial Activities of Catechol-Conjugated β-Lactams against Gram-Negative Bacteria

Cefiderocol is the first approved catechol-conjugated cephalosporin against multidrug-resistant Gram-negative bacteria, while its application was limited by poor chemical stability associated with the pyrrolidinium linker, moderate potency against Klebsiella pneumoniae and Acinetobacter baumannii, intricate procedures for salt preparation, and potential hypersensitivity. To address these issues, a series of novel catechol-conjugated derivatives were designed, synthesized, and evaluated. Extensive structure-activity relationships and structure-metabolism relationships (SMR) were conducted, leading to the discovery of a promising compound 86b (Code no. YFJ-36) with a new thioether linker. 86b exhibited superior and broad-spectrum in vitro antibacterial activity, especially against A. baumannii and K. pneumoniae, compared with cefiderocol. Potent in vivo efficacy was observed in a murine systemic infection model. Furthermore, the physicochemical stability of 86b in fluid medium at pH 6-8 was enhanced. 86b also reduced potential the risk of allergy owing to the quaternary ammonium linker. The improved properties of 86b supported its further research and development.

Discovery of Highly Potent Solute Carrier 13 Member 5 (SLC13A5) Inhibitors for the Treatment of Hyperlipidemia

In the face of escalating metabolic disease prevalence, largely driven by modern lifestyle factors, this study addresses the critical need for novel therapeutic approaches. We have identified the sodium-coupled citrate transporter (NaCT or SLC13A5) as a target for intervention. Utilizing rational drug design, we developed a new class of SLC13A5 inhibitors, anchored by the hydroxysuccinic acid scaffold, refining the structure of PF-06649298. Among these, LBA-3 emerged as a standout compound, exhibiting remarkable potency with an IC50 value of 67 nM, significantly improving upon PF-06649298. In vitro assays demonstrated LBA-3's efficacy in reducing triglyceride levels in OPA-induced HepG2 cells. Moreover, LBA-3 displayed superior pharmacokinetic properties and effectively lowered triglyceride and total cholesterol levels in diverse mouse models (PCN-stimulated and starvation-induced), without detectable toxicity. These findings not only spotlight LBA-3 as a promising candidate for hyperlipidemia treatment but also exemplify the potential of targeted molecular design in advancing metabolic disorder therapeutics.

Discovery of TNG908: A Selective, Brain Penetrant, MTA-Cooperative PRMT5 Inhibitor That Is Synthetically Lethal with MTAP-Deleted Cancers

It has been shown that PRMT5 inhibition by small molecules can selectively kill cancer cells with homozygous deletion of the MTAP gene if the inhibitors can leverage the consequence of MTAP deletion, namely, accumulation of the MTAP substrate MTA. Herein, we describe the discovery of TNG908, a potent inhibitor that binds the PRMT5·MTA complex, leading to 15-fold-selective killing of MTAP-deleted (MTAP-null) cells compared to MTAPintact (MTAP WT) cells. TNG908 shows selective antitumor activity when dosed orally in mouse xenograft models, and its physicochemical properties are amenable for crossing the blood-brain barrier (BBB), supporting clinical study for the treatment of both CNS and non-CNS tumors with MTAP loss.

Discovery of Novel Aryl Triazolone Dihydropyridines (ATDPs) Targeting Highly Conserved Residue W229 as Promising HIV-1 NNRTIs

NNRTI is an important component of the highly active antiretroviral therapy (HAART), but the rapid emergence of drug resistance and poor pharmacokinetics limited their clinical application. Herein, a series of novel aryl triazolone dihydropyridines (ATDPs) were designed by structure-guided design with the aim of improving drug resistance profiles and pharmacokinetic profiles. Compound 10n (EC50 = 0.009-17.7 μM) exhibited the most active potency, being superior to or comparable to that of doravirine (DOR) against the whole tested viral panel. Molecular docking was performed to clarify the reason for its higher resistance profiles. Moreover, 10n demonstrated excellent pharmacokinetic profile (T1/2 = 5.09 h, F = 108.96%) compared that of DOR (T1/2 = 4.4 h, F = 57%). Additionally, 10n was also verified to have no in vivo acute or subacute toxicity (LD50 > 2000 mg/kg), suggesting that 10n is worth further investigation as a novel oral NNRTIs for HIV-1 therapy.

Synthesis and Structure-Activity Relationships of 2,5-Dimethoxy-4-Substituted Phenethylamines and the Discovery of CYB210010: A Potent, Orally Bioavailable and Long-Acting Serotonin 5-HT2 Receptor Agonist

Structure-activity studies of 4-substituted-2,5-dimethoxyphenethylamines led to the discovery of 2,5-dimethoxy-4-thiotrifluoromethylphenethylamines, including CYB210010, a potent and long-acting serotonin 5-HT2 receptor agonist. CYB210010 exhibited high agonist potency at 5-HT2A and 5-HT2C receptors, modest selectivity over 5-HT2B, 5-HT1A, 5-HT6, and adrenergic α2A receptors, and lacked activity at monoamine transporters and over 70 other proteins. CYB210010 (0.1-3 mg/kg) elicited a head-twitch response (HTR) and could be administered subchronically at threshold doses without behavioral tolerance. CYB210010 was orally bioavailable in three species, readily and preferentially crossed into the CNS, engaged frontal cortex 5-HT2A receptors, and increased the expression of genes involved in neuroplasticity in the frontal cortex. CYB210010 represents a new tool molecule for investigating the therapeutic potential of 5-HT2 receptor activation. In addition, several other compounds with high 5-HT2A receptor potency, yet with little or no HTR activity, were discovered, providing the groundwork for the development of nonpsychedelic 5-HT2A receptor ligands.

Discovery of a Novel Potent Tetrazole Antifungal Candidate with High Selectivity and Broad Spectrum

Thirty-one novel albaconazole derivatives were designed and synthesized based on our previous work. All compounds exhibited potent in vitro antifungal activities against seven pathogenic fungi. Among them, tetrazole compound D2 was the most potent antifungal with MIC values of <0.008, <0.008, and 2 μg/mL against Candida albicans, Cryptococcus neoformans, and Aspergillus fumigatus, respectively, the three most common and critical priority pathogenic fungi. In addition, compound D2 also exhibited potent activity against fluconazole-resistant C. auris isolates. Notably, compound D2 showed a lower inhibitory activity in vitro against human CYP450 enzymes as well as a lower inhibitory effect on the hERG K+ channel, indicating a low risk of drug-drug interactions and QT prolongation. Moreover, with improved pharmacokinetic profiles, compound D2 showed better in vivo efficacy than albaconazole at reducing fungal burden and extending the survival of C. albicans-infected mice. Taken together, compound D2 will be further investigated as a promising candidate.

Discovery of First-in-Class PROTAC Degraders of SARS-CoV-2 Main Protease

We have witnessed three coronavirus (CoV) outbreaks in the past two decades, including the COVID-19 pandemic caused by SARS-CoV-2. Main protease (MPro), a highly conserved protease among various CoVs, is essential for viral replication and pathogenesis, making it a prime target for antiviral drug development. Here, we leverage proteolysis targeting chimera (PROTAC) technology to develop a new class of small-molecule antivirals that induce the degradation of SARS-CoV-2 MPro. Among them, MPD2 was demonstrated to effectively reduce MPro protein levels in 293T cells, relying on a time-dependent, CRBN-mediated, and proteasome-driven mechanism. Furthermore, MPD2 exhibited remarkable efficacy in diminishing MPro protein levels in SARS-CoV-2-infected A549-ACE2 cells. MPD2 also displayed potent antiviral activity against various SARS-CoV-2 strains and exhibited enhanced potency against nirmatrelvir-resistant viruses. Overall, this proof-of-concept study highlights the potential of targeted protein degradation of MPro as an innovative approach for developing antivirals that could fight against drug-resistant viral variants.

Discovery of Small Molecule Interleukin 17A Inhibitors with Novel Binding Mode and Stoichiometry: Optimization of DNA-Encoded Chemical Library Hits to In Vivo Active Compounds

Dysregulation of IL17A drives numerous inflammatory and autoimmune disorders with inhibition of IL17A using antibodies proven as an effective treatment. Oral anti-IL17 therapies are an attractive alternative option, and several preclinical small molecule IL17 inhibitors have previously been described. Herein, we report the discovery of a novel class of small molecule IL17A inhibitors, identified via a DNA-encoded chemical library screen, and their subsequent optimization to provide in vivo efficacious inhibitors. These new protein-protein interaction (PPI) inhibitors bind in a previously undescribed mode in the IL17A protein with two copies binding symmetrically to the central cavities of the IL17A homodimer.

Pharmacokinetics Profiler (PhaKinPro): Model Development, Validation, and Implementation as a Web Tool for Triaging Compounds with Undesired Pharmacokinetics Profiles

Computational models that predict pharmacokinetic properties are critical to deprioritize drug candidates that emerge as hits in high-throughput screening campaigns. We collected, curated, and integrated a database of compounds tested in 12 major end points comprising over 10,000 unique molecules. We then employed these data to build and validate binary quantitative structure-activity relationship (QSAR) models. All trained models achieved a correct classification rate above 0.60 and a positive predictive value above 0.50. To illustrate their utility in drug discovery, we used these models to predict the pharmacokinetic properties for drugs in the NCATS Inxight Drugs database. In addition, we employed the developed models to predict the pharmacokinetic properties of all compounds in the DrugBank. All models described in this paper have been integrated and made publicly available via the PhaKinPro Web-portal that can be accessed at https://phakinpro.mml.unc.edu/.

Discovery of Covalent Lead Compounds Targeting 3CL Protease with a Lateral Interactions Spiking Neural Network

Covalent drugs exhibit advantages in that noncovalent drugs cannot match, and covalent docking is an important method for screening covalent lead compounds. However, it is difficult for covalent docking to screen covalent compounds on a large scale because covalent docking requires determination of the covalent reaction type of the compound. Here, we propose to use deep learning of a lateral interactions spiking neural network to construct a covalent lead compound screening model to quickly screen covalent lead compounds. We used the 3CL protease (3CL Pro) of SARS-CoV-2 as the screen target and constructed two classification models based on LISNN to predict the covalent binding and inhibitory activity of compounds. The two classification models were trained on the covalent complex data set targeting cysteine (Cys) and the compound inhibitory activity data set targeting 3CL Pro, respected, with good prediction accuracy (ACC > 0.9). We then screened the screening compound library with 6 covalent binding screening models and 12 inhibitory activity screening models. We tested the inhibitory activity of the 32 compounds, and the best compound inhibited SARS-CoV-2 3CL Pro with an IC50 value of 369.5 nM. Further assay implied that dithiothreitol can affect the inhibitory activity of the compound to 3CL Pro, indicating that the compound may covalently bind 3CL Pro. The selectivity test showed that the compound had good target selectivity to 3CL Pro over cathepsin L. These correlation assays can prove the rationality of the covalent lead compound screening model. Finally, covalent docking was performed to demonstrate the binding conformation of the compound with 3CL Pro. The source code can be obtained from the GitHub repository (https://github.com/guzh970630/Screen_Covalent_Compound_by_LISNN).

Protein characteristics substantially influence the propensity of activity cliffs among kinase inhibitors

Activity cliffs (ACs) are pairs of structurally similar molecules with significantly different affinities for a biotarget, posing a challenge in computer-assisted drug discovery. This study focuses on protein kinases, significant therapeutic targets, with some exhibiting ACs while others do not despite numerous inhibitors. The hypothesis that the presence of ACs is dependent on the target protein and its complete structural context is explored. Machine learning models were developed to link protein properties to ACs, revealing specific tripeptide sequences and overall protein properties as critical factors in ACs occurrence. The study highlights the importance of considering the entire protein matrix rather than just the binding site in understanding ACs. This research provides valuable insights for drug discovery and design, paving the way for addressing ACs-related challenges in modern computational approaches.

A comprehensive dataset of protein-protein interactions and ligand binding pockets for advancing drug discovery

This dataset represents a collection of pocket-centric structural data related to protein-protein interactions (PPIs) and PPI-related ligand binding sites. The dataset includes high-quality structural information on more than 23,000 pockets, 3,700 proteins on more than 500 organisms, and nearly 3500 ligands that can aid researchers in the fields of bioinformatics, structural biology, and drug discovery. It encompasses a diverse set of PPI complexes with more than 1,700 unique protein families including some with associated ligands, enabling detailed investigations into molecular interactions at the atomic level. This article introduces an indispensable resource designed to unlock the full potential of PPIs while pioneering a novel metric for pocket similarity for hypothesizing protein partners repurposing.

Discovery of a novel and highly selective JAK3 inhibitor as a potent hair growth promoter

JAK-STAT signalling pathway inhibitors have emerged as promising therapeutic agents for the treatment of hair loss. Among different JAK isoforms, JAK3 has become an ideal target for drug discovery because it only regulates a narrow spectrum of γc cytokines. Here, we report the discovery of MJ04, a novel and highly selective 3-pyrimidinylazaindole based JAK3 inhibitor, as a potential hair growth promoter with an IC50 of 2.03 nM. During in vivo efficacy assays, topical application of MJ04 on DHT-challenged AGA and athymic nude mice resulted in early onset of hair regrowth. Furthermore, MJ04 significantly promoted the growth of human hair follicles under ex-vivo conditions. MJ04 exhibited a reasonably good pharmacokinetic profile and demonstrated a favourable safety profile under in vivo and in vitro conditions. Taken together, we report MJ04 as a highly potent and selective JAK3 inhibitor that exhibits overall properties suitable for topical drug development and advancement to human clinical trials.

Drug Design: Do Not Forget the Supramolecular Factor

A major challenge currently facing medicinal chemists is designing agents that can selectively destroy drug resistant fungi and bacteria that have begun to emerge. One factor that has been overlooked by virtually all drug discovery/development approaches is the supramolecular factor, in which aggregated forms of a drug candidate exhibit low selectivity in destroying targeted cells while the corresponding monomers exhibit high selectivity. This Perspective discusses how we were led to the supramolecular factor through fundamental studies with simple model systems, how we reasoned that the selectivity of monomers of the antifungal agent amphotericin B should be much greater than the selectivity of the corresponding aggregates, and how we confirmed this hypothesis using derivatives of amphotericin B. In a broader context, these findings provide a strong rationale for considering the supramolecular factor in the design of new drug candidates and the testing of virtually all of them.

Performance and robustness of small molecule retention time prediction with molecular graph neural networks in industrial drug discovery campaigns

This study explores how machine-learning can be used to predict chromatographic retention times (RT) for the analysis of small molecules, with the objective of identifying a machine-learning framework with the robustness required to support a chemical synthesis production platform. We used internally generated data from high-throughput parallel synthesis in context of pharmaceutical drug discovery projects. We tested machine-learning models from the following frameworks: XGBoost, ChemProp, and DeepChem, using a dataset of 7552 small molecules. Our findings show that two specific models, AttentiveFP and ChemProp, performed better than XGBoost and a regular neural network in predicting RT accurately. We also assessed how well these models performed over time and found that molecular graph neural networks consistently gave accurate predictions for new chemical series. In addition, when we applied ChemProp on the publicly available METLIN SMRT dataset, it performed impressively with an average error of 38.70 s. These results highlight the efficacy of molecular graph neural networks, especially ChemProp, in diverse RT prediction scenarios, thereby enhancing the efficiency of chromatographic analysis.

Bioisoteres for carboxylic acids: From ionized isosteres to novel unionized replacements

Carboxylic acids are key pharmacophoric elements in many molecules. They can be seen as a problem by some, due to perceived permeability challenges, potential for high plasma protein binding and the risk of forming reactive metabolites due to acyl-glucuronidation. By others they are viewed more favorably as they can decrease lipophilicity by adding an ionizable center which can be beneficial for solubility, and can add enthalpic interactions with the target protein. However, there are many instances where the replacement of a carboxylic acid with a bioisosteric group is required. This has led to the development of a number of ionizable groups which sufficiently mimic the carboxylic acid functionality whilst improving, for example, the metabolic profile of the molecule in question. An alternative strategy involves replacement of the carboxylate by neutral functional groups. This review initially details carefully selected examples whereby tetrazoles, acyl sulfonamides or isoxazolols have been beneficially utilized as carboxylic acid bioisosteres altering physicohemical properties, interactions with the target and metabolism and/or pharmacokinetics, before delving further into the binding mode of carboxylic acid derivatives with their target proteins. This analysis highlights new ways to consider the replacement of carboxylic acids by neutral bioisosteric groups which either rely on hydrogen bonds or cation-π interactions. It should serve as a useful guide for scientists working in drug discovery.

G protein-coupled receptors (GPCRs): advances in structures, mechanisms, and drug discovery

G protein-coupled receptors (GPCRs), the largest family of human membrane proteins and an important class of drug targets, play a role in maintaining numerous physiological processes. Agonist or antagonist, orthosteric effects or allosteric effects, and biased signaling or balanced signaling, characterize the complexity of GPCR dynamic features. In this study, we first review the structural advancements, activation mechanisms, and functional diversity of GPCRs. We then focus on GPCR drug discovery by revealing the detailed drug-target interactions and the underlying mechanisms of orthosteric drugs approved by the US Food and Drug Administration in the past five years. Particularly, an up-to-date analysis is performed on available GPCR structures complexed with synthetic small-molecule allosteric modulators to elucidate key receptor-ligand interactions and allosteric mechanisms. Finally, we highlight how the widespread GPCR-druggable allosteric sites can guide structure- or mechanism-based drug design and propose prospects of designing bitopic ligands for the future therapeutic potential of targeting this receptor family.

Boosting Clear Cell Renal Carcinoma-Specific Drug Discovery Using a Deep Learning Algorithm and Single-Cell Analysis

Clear cell renal carcinoma (ccRCC), the most common subtype of renal cell carcinoma, has the high heterogeneity of a highly complex tumor microenvironment. Existing clinical intervention strategies, such as target therapy and immunotherapy, have failed to achieve good therapeutic effects. In this article, single-cell transcriptome sequencing (scRNA-seq) data from six patients downloaded from the GEO database were adopted to describe the tumor microenvironment (TME) of ccRCC, including its T cells, tumor-associated macrophages (TAMs), endothelial cells (ECs), and cancer-associated fibroblasts (CAFs). Based on the differential typing of the TME, we identified tumor cell-specific regulatory programs that are mediated by three key transcription factors (TFs), whilst the TF EPAS1/HIF-2α was identified via drug virtual screening through our analysis of ccRCC's protein structure. Then, a combined deep graph neural network and machine learning algorithm were used to select anti-ccRCC compounds from bioactive compound libraries, including the FDA-approved drug library, natural product library, and human endogenous metabolite compound library. Finally, five compounds were obtained, including two FDA-approved drugs (flufenamic acid and fludarabine), one endogenous metabolite, one immunology/inflammation-related compound, and one inhibitor of DNA methyltransferase (N4-methylcytidine, a cytosine nucleoside analogue that, like zebularine, has the mechanism of inhibiting DNA methyltransferase). Based on the tumor microenvironment characteristics of ccRCC, five ccRCC-specific compounds were identified, which would give direction of the clinical treatment for ccRCC patients.

SerotoninAI: Serotonergic System Focused, Artificial Intelligence-Based Application for Drug Discovery

SerotoninAI is an innovative web application for scientific purposes focused on the serotonergic system. By leveraging SerotoninAI, researchers can assess the affinity (pKi value) of a molecule to all main serotonin receptors and serotonin transporters based on molecule structure introduced as SMILES. Additionally, the application provides essential insights into critical attributes of potential drugs such as blood-brain barrier penetration and human intestinal absorption. The complexity of the serotonergic system demands advanced tools for accurate predictions, which is a fundamental requirement in drug development. SerotoninAI addresses this need by providing an intuitive user interface that generates predictions of pKi values for the main serotonergic targets. The application is freely available on the Internet at https://serotoninai.streamlit.app/, implemented in Streamlit with all major web browsers supported. Currently, to the best of our knowledge, there is no tool that allows users to access affinity predictions for serotonergic targets without registration or financial obligations. SerotoninAI significantly increases the scope of drug development activities worldwide. The source code of the application is available at https://github.com/nczub/SerotoninAI_streamlit.

Graph Attention Site Prediction (GrASP): Identifying Druggable Binding Sites Using Graph Neural Networks with Attention

Identifying and discovering druggable protein binding sites is an important early step in computer-aided drug discovery, but it remains a difficult task where most campaigns rely on a priori knowledge of binding sites from experiments. Here, we present a binding site prediction method called Graph Attention Site Prediction (GrASP) and re-evaluate assumptions in nearly every step in the site prediction workflow from data set preparation to model evaluation. GrASP is able to achieve state-of-the-art performance at recovering binding sites in PDB structures while maintaining a high degree of precision which will minimize wasted computation in downstream tasks such as docking and free energy perturbation.

HyperPCM: Robust Task-Conditioned Modeling of Drug-Target Interactions

A central problem in drug discovery is to identify the interactions between drug-like compounds and protein targets. Over the past few decades, various quantitative structure-activity relationship (QSAR) and proteo-chemometric (PCM) approaches have been developed to model and predict these interactions. While QSAR approaches solely utilize representations of the drug compound, PCM methods incorporate both representations of the protein target and the drug compound, enabling them to achieve above-chance predictive accuracy on previously unseen protein targets. Both QSAR and PCM approaches have recently been improved by machine learning and deep neural networks, that allow the development of drug-target interaction prediction models from measurement data. However, deep neural networks typically require large amounts of training data and cannot robustly adapt to new tasks, such as predicting interaction for unseen protein targets at inference time. In this work, we propose to use HyperNetworks to efficiently transfer information between tasks during inference and thus to accurately predict drug-target interactions on unseen protein targets. Our HyperPCM method reaches state-of-the-art performance compared to previous methods on multiple well-known benchmarks, including Davis, DUD-E, and a ChEMBL derived data set, and particularly excels at zero-shot inference involving unseen protein targets. Our method, as well as reproducible data preparation, is available at https://github.com/ml-jku/hyper-dti.

Deep Generative Models in De Novo Drug Molecule Generation

The discovery of new drugs has important implications for human health. Traditional methods for drug discovery rely on experiments to optimize the structure of lead molecules, which are time-consuming and high-cost. Recently, artificial intelligence has exhibited promising and efficient performance for drug-like molecule generation. In particular, deep generative models achieve great success in de novo generation of drug-like molecules with desired properties, showing massive potential for novel drug discovery. In this study, we review the recent progress of molecule generation using deep generative models, mainly focusing on molecule representations, public databases, data processing tools, and advanced artificial intelligence based molecule generation frameworks. In particular, we present a comprehensive comparison of state-of-the-art deep generative models for molecule generation and a summary of commonly used molecular design strategies. We identify research gaps and challenges of molecule generation such as the need for better databases, missing 3D information in molecular representation, and the lack of high-precision evaluation metrics. We suggest future directions for molecular generation and drug discovery.

Lysosomal storage, impaired autophagy and innate immunity in Gaucher and Parkinson's diseases: insights for drug discovery

Impairment of autophagic-lysosomal pathways is increasingly being implicated in Parkinson's disease (PD). GBA1 mutations cause the lysosomal storage disorder Gaucher disease (GD) and are the commonest known genetic risk factor for PD. GBA1 mutations have been shown to cause autophagic-lysosomal impairment. Defective autophagic degradation of unwanted cellular constituents is associated with several pathologies, including loss of normal protein homeostasis, particularly of α-synuclein, and innate immune dysfunction. The latter is observed both peripherally and centrally in PD and GD. Here, we will discuss the mechanistic links between autophagy and immune dysregulation, and the possible role of these pathologies in communication between the gut and brain in these disorders. Recent work in a fly model of neuronopathic GD (nGD) revealed intestinal autophagic defects leading to gastrointestinal dysfunction and immune activation. Rapamycin treatment partially reversed the autophagic block and reduced immune activity, in association with increased survival and improved locomotor performance. Alterations in the gut microbiome are a critical driver of neuroinflammation, and studies have revealed that eradication of the microbiome in nGD fly and mouse models of PD ameliorate brain inflammation. Following these observations, lysosomal-autophagic pathways, innate immune signalling and microbiome dysbiosis are discussed as potential therapeutic targets in PD and GD. This article is part of a discussion meeting issue 'Understanding the endo-lysosomal network in neurodegeneration'.

DFRscore: Deep Learning-Based Scoring of Synthetic Complexity with Drug-Focused Retrosynthetic Analysis for High-Throughput Virtual Screening

Recently emerging generative AI models enable us to produce a vast number of compounds for potential applications. While they can provide novel molecular structures, the synthetic feasibility of the generated molecules is often questioned. To address this issue, a few recent studies have attempted to use deep learning models to estimate the synthetic accessibility of many molecules rapidly. However, retrosynthetic analysis tools used to train the models rely on reaction templates automatically extracted from a large reaction database that are not domain-specific and may exhibit low chemical correctness. To overcome this limitation, we introduce DFRscore (Drug-Focused Retrosynthetic score), a deep learning-based approach for a more practical assessment of synthetic accessibility in drug discovery. The DFRscore model is trained exclusively on drug-focused reactions, providing a predicted number of minimally required synthetic steps for each compound. This approach enables practitioners to filter out compounds that do not meet their desired level of synthetic accessibility at an early stage of high-throughput virtual screening for accelerated drug discovery. The proposed strategy can be easily adapted to other domains by adjusting the synthesis planning setup of the reaction templates and starting materials.

Regression-Based Active Learning for Accessible Acceleration of Ultra-Large Library Docking

Structure-based drug discovery is a process for both hit finding and optimization that relies on a validated three-dimensional model of a target biomolecule, used to rationalize the structure-function relationship for this particular target. An ultralarge virtual screening approach has emerged recently for rapid discovery of high-affinity hit compounds, but it requires substantial computational resources. This study shows that active learning with simple linear regression models can accelerate virtual screening, retrieving up to 90% of the top-1% of the docking hit list after docking just 10% of the ligands. The results demonstrate that it is unnecessary to use complex models, such as deep learning approaches, to predict the imprecise results of ligand docking with a low sampling depth. Furthermore, we explore active learning meta-parameters and find that constant batch size models with a simple ensembling method provide the best ligand retrieval rate. Finally, our approach is validated on the ultralarge size virtual screening data set, retrieving 70% of the top-0.05% of ligands after screening only 2% of the library. Altogether, this work provides a computationally accessible approach for accelerated virtual screening that can serve as a blueprint for the future design of low-compute agents for exploration of the chemical space via large-scale accelerated docking. With recent breakthroughs in protein structure prediction, this method can significantly increase accessibility for the academic community and aid in the rapid discovery of high-affinity hit compounds for various targets.

graphLambda: Fusion Graph Neural Networks for Binding Affinity Prediction

Predicting the binding affinity of protein-ligand complexes is crucial for computer-aided drug discovery (CADD) and the identification of potential drug candidates. The deep learning-based scoring functions have emerged as promising predictors of binding constants. Building on recent advancements in graph neural networks, we present graphLambda for protein-ligand binding affinity prediction, which utilizes graph convolutional, attention, and isomorphism blocks to enhance the predictive capabilities. The graphLambda model exhibits superior performance across CASF16 and CSAR HiQ NRC benchmarks and demonstrates robustness with respect to different types of train-validation set partitions. The development of graphLambda underscores the potential of graph neural networks in advancing binding affinity prediction models, contributing to more effective CADD methodologies.

Explainable Graph Neural Networks with Data Augmentation for Predicting pKa of C-H Acids

The pKa of C-H acids is an important parameter in the fields of organic synthesis, drug discovery, and materials science. However, the prediction of pKa is still a great challenge due to the limit of experimental data and the lack of chemical insight. Here, a new model for predicting the pKa values of C-H acids is proposed on the basis of graph neural networks (GNNs) and data augmentation. A message passing unit (MPU) was used to extract the topological and target-related information from the molecular graph data, and a readout layer was utilized to retrieve the information on the ionization site C atom. The retrieved information then was adopted to predict pKa by a fully connected network. Furthermore, to increase the diversity of the training data, a knowledge-infused data augmentation technique was established by replacing the H atoms in a molecule with substituents exhibiting different electronic effects. The MPU was pretrained with the augmented data. The efficacy of data augmentation was confirmed by visualizing the distribution of compounds with different substituents and by classifying compounds. The explainability of the model was studied by examining the change of pKa values when a specific atom was masked. This explainability was used to identify the key substituents for pKa. The model was evaluated on two data sets from the iBonD database. Dataset1 includes the experimental pKa values of C-H acids measured in DMSO, while dataset2 comprises the pKa values measured in water. The results show that the knowledge-infused data augmentation technique greatly improves the predictive accuracy of the model, especially when the number of samples is small.

Can Pretrained Models Really Learn Better Molecular Representations for AI-Aided Drug Discovery?

Self-supervised pretrained models are gaining increasingly more popularity in AI-aided drug discovery, leading to more and more pretrained models with the promise that they can extract better feature representations for molecules. Yet, the quality of learned representations has not been fully explored. In this work, inspired by the two phenomena of Activity Cliffs (ACs) and Scaffold Hopping (SH) in traditional Quantitative Structure-Activity Relationship analysis, we propose a method named Representation-Property Relationship Analysis (RePRA) to evaluate the quality of the representations extracted by the pretrained model and visualize the relationship between the representations and properties. The concepts of ACs and SH are generalized from the structure-activity context to the representation-property context, and the underlying principles of RePRA are analyzed theoretically. Two scores are designed to measure the generalized ACs and SH detected by RePRA, and therefore, the quality of representations can be evaluated. In experiments, representations of molecules from 10 target tasks generated by 7 pretrained models are analyzed. The results indicate that the state-of-the-art pretrained models can overcome some shortcomings of canonical Extended-Connectivity FingerPrints, while the correlation between the basis of the representation space and specific molecular substructures are not explicit. Thus, some representations could be even worse than the canonical fingerprints. Our method enables researchers to evaluate the quality of molecular representations generated by their proposed self-supervised pretrained models. And our findings can guide the community to develop better pretraining techniques to regularize the occurrence of ACs and SH.

Streamlining Large Chemical Library Docking with Artificial Intelligence: the PyRMD2Dock Approach

The present contribution introduces a novel computational protocol called PyRMD2Dock, which combines the Ligand-Based Virtual Screening (LBVS) tool PyRMD with the popular docking software AutoDock-GPU (AD4-GPU) to enhance the throughput of virtual screening campaigns for drug discovery. By implementing PyRMD2Dock, we demonstrate that it is possible to rapidly screen massive chemical databases and identify those with the highest predicted binding affinity to a target protein. Our benchmarking and screening experiments illustrate the predictive power and speed of PyRMD2Dock and highlight its potential to accelerate the discovery of novel drug candidates. Overall, this study showcases the value of combining AI-powered LBVS tools with docking software to enable effective and high-throughput virtual screening of ultralarge molecular databases in drug discovery. PyRMD and the PyRMD2Dock protocol are freely available on GitHub (https://github.com/cosconatilab/PyRMD) as an open-source tool.