Identification of novel LRRK2 inhibitors by structure-based virtual screening and alchemical free energy calculation.

The Leucine-rich repeat kinase 2 (LRRK2) target has been identified as a promising drug target for Parkinson's disease (PD) treatment. This study focuses on optimizing the activity of LRRK2 inhibitors using alchemical relative binding free energy (RBFE) calculations. Initially, we assessed various free energy calculation methods across different LRRK2 kinase inhibitor scaffolds. The results indicate that alchemical free energy calculations are promising for prospective predictions on LRRK2 inhibitors, especially for the aminopyrimidine scaffold with an RMSE of 1.15 kcal mol-1 and Rp of 0.83. Following this, we optimized a potent LRRK2 kinase inhibitor identified from previous virtual screenings, featuring a novel scaffold. Guided by RBFE predictions using alchemical methods, this optimization led to the discovery of compound LY2023-001. This compound, with a [1,2,4]triazolo[5,6-b]indole scaffold, exhibited enhanced inhibitory activity against G2019S LRRK2 (IC50 = 12.9 nM). Molecular dynamics (MD) simulations revealed that LY2023-001 formed stable hydrogen bonds with Glu1948, and Ala1950 in the G2019S LRRK2 protein. Additionally, its phenyl substituents engage in strong electrostatic interactions with Lys1906 and van der Waals interactions with Leu1885, Phe1890, Val1893, Ile1933, Met1947, Leu1949, Leu2001, Ala2016, and Asp2017. Our findings underscore the potential of computational methods in the successful optimization of small molecules, offering important insights for the development of novel LRRK2 inhibitors.

Versatile Framework for Drug-Target Interaction Prediction by Considering Domain-Specific Features.

Predicting drug-target interactions (DTIs) is one of the crucial tasks in drug discovery, but traditional wet-lab experiments are costly and time-consuming. Recently, deep learning has emerged as a promising tool for accelerating DTI prediction due to its powerful performance. However, the models trained on limited known DTI data struggle to generalize effectively to novel drug-target pairs. In this work, we propose a strategy to train an ensemble of models by capturing both domain-generic and domain-specific features (E-DIS) to learn diverse domain features and adapt them to out-of-distribution data. Multiple experts were trained on different domains to capture and align domain-specific information from various distributions without accessing any data from unseen domains. E-DIS provides a comprehensive representation of proteins and ligands by capturing diverse features. Experimental results on four benchmark data sets in both in-domain and cross-domain settings demonstrated that E-DIS significantly improved model performance and domain generalization compared to existing methods. Our approach presents a significant advancement in DTI prediction by combining domain-generic and domain-specific features, enhancing the generalization ability of the DTI prediction model.

Synthesis and Biological Activity of 2-Chloro-8-methoxy-5-methyl-5H-indolo [2,3-b] Quinoline for the Treatment of Colorectal Cancer by Modulating PI3K/AKT/mTOR Pathways.

Developing novel drugs from natural products has proven to be a very effective strategy. Neocryptolepine was isolated from Cryptolepis sanguinolenta, a traditional endemic African herb, which exerts a wide range of biological activities such as antimalaria, antibacterial, and antitumor. 2-Chloro-8-methoxy-5-methyl-5H-indolo [2,3-b] quinoline (compound 49) was synthesized, and its cytotoxicity was assessed on pancreatic cancer PANC-1 cells, colorectal cancer HCT116 cells, liver cancer SMMC-7721 cells, and gastric cancer AGS cells in vitro. The results of the in vitro assay showed that compound 49 exerted remarkable cytotoxicity on colorectal cancer HCT116 and Caco-2 cells. The cytotoxicity of compound 49 to colorectal cancer HCT116 cells was 17 times higher than that of neocryptolepine and to human normal intestinal epithelial HIEC cells was significantly reduced. Compound 49 exhibited significant cytotoxicity against the colorectal cancer HCT116 and Caco-2 cells, with IC50 of 0.35 and 0.54 µM, respectively. The mechanism of cytotoxicity of compound 49 to colorectal cancer HCT116 and Caco-2 cells was further investigated. The results showed that compound 49 could inhibit colony formation and cell migration. Moreover, compound 49 could arrest the cell cycle at the G2/M phase, promote the production of reactive oxygen species, reduce mitochondrial membrane potential, and induce apoptosis. The results of Western blot indicated that compound 49 showed cytotoxicity on HCT116 and Caco-2 cells by modulating the PI3K/AKT/mTOR signaling pathway. In conclusion, these results suggested that compound 49 may be a potentially promising lead compound for the treatment of colorectal cancer.

Strategic advancement of E3 ubiquitin ligase in the management of hepatocellular carcinoma.

Hepatocellular carcinoma (HCC) presents a significant global health challenge due to its high incidence, poor prognosis, and limited treatment options. As a pivotal regulator of protein stability, E3 ubiquitin ligase plays a crucial role in tumorigenesis and development. This review provides an overview of the latest research on the involvement of E3 ubiquitin ligase in hepatocellular carcinoma and elucidates its significance in hepatocellular carcinoma cell proliferation, invasion, and evasion from immune surveillance. Special attention is given to the functions of RING, HECT, and RBR E3 ubiquitin ligases and their association with hepatocellular carcinoma progression. By dissecting the molecular mechanisms and regulatory networks governed by E3 ubiquitin ligase, several potential therapeutic strategies are proposed: including the development of specific inhibitors targeting E3 ligases; augmentation of their tumor suppressor activity through drug or gene therapy; utilization of E3 ubiquitin ligase to modulate immune checkpoint proteins for improved efficacy of immunotherapy; combination strategies integrating traditional therapies with E3 ubiquitin ligase inhibitors; as well as biomarker development based on E3 ubiquitin ligase activity. Furthermore, this review discusses the prospect of overcoming drug resistance in hepatocellular carcinoma treatment through these novel approaches. Overall, this review establishes a theoretical foundation and offers fresh insights into harnessing the potential of E3 ubiquitin ligase for treating hepatocellular carcinoma while highlighting future research directions that pave the way for clinical translation studies and new drug discoveries.

Versatile Self-Assembly of Triblock Peptides into Stable Collagen Mimetic Heterotrimers.

The construction of peptides to mimic heterogeneous proteins such as type I collagen plays a pivotal role in deciphering their function and pathogenesis. However, progress in the field has been severely hampered by the lack of capability to create stable heterotrimers with desired functional sequences and without the effect of homotrimers. We have herein developed a set of triblock peptides that can assemble into collagen mimetic heterotrimers with desired amino acids and are free from the interference of homotrimers. The triblock peptides comprise a central collagen-like block and two oppositely charged N-/C-terminal blocks, which display inherent incompetency of homotrimer formation. The favorable electrostatic attraction between two paired triblock peptides with complementary terminal charged sequences promptly leads to stable heterotrimers with controlled chain composition. The independence of the collagen-like block from the two terminal blocks endows this system with the adaptability to incorporate desired amino acid sequences while maintaining the heterotrimer structure. The triblock peptides provide a versatile and robust tool to mimic the composition and function of heterotrimer collagen and may have great potential in the design of innovative peptides mimicking heterogeneous proteins.

Multiscale calculations reveal new insights into the reaction mechanism between KRASG12C and α, ß-unsaturated carbonyl of covalent inhibitors.

Utilizing α,ß-unsaturated carbonyl group as Michael acceptors to react with thiols represents a successful strategy for developing KRASG12C inhibitors. Despite this, the precise reaction mechanism between KRASG12C and covalent inhibitors remains a subject of debate, primarily due to the absence of an appropriate residue capable of deprotonating the cysteine thiol as a base. To uncover this reaction mechanism, we first discussed the chemical reaction mechanism in solvent conditions via density functional theory (DFT) calculation. Based on this, we then proposed and validated the enzymatic reaction mechanism by employing quantum mechanics/molecular mechanics (QM/MM) calculation. Our QM/MM analysis suggests that, in biological conditions, proton transfer and nucleophilic addition may proceed through a concerted process to form an enolate intermediate, bypassing the need for a base catalyst. This proposed mechanism differs from previous findings. Following the formation of the enolate intermediate, solvent-assisted tautomerization results in the final product. Our calculations indicate that solvent-assisted tautomerization is the rate-limiting step in the catalytic cycle under biological conditions. On the basis of this reaction mechanism, the calculated kinact/ki for two inhibitors is consistent well with the experimental results. Our findings provide new insights into the reaction mechanism between the cysteine of KRASG12C and the covalent inhibitors and may provide valuable information for designing effective covalent inhibitors targeting KRASG12C and other similar targets.

Study of andrographolide bioactivity against Pseudomonas aeruginosa based on computational methodology and biochemical analysis.

Andrographolide is one of the main biologically active molecules isolated from Andrographis paniculata (A. paniculata), which is a traditional Chinese herb used extensively throughout Eastern Asia, India, and China. Pseudomonas aeruginosa, often known as P. aeruginosa, is a common clinical opportunistic pathogen with remarkable adaptability to harsh settings and resistance to antibiotics. P. aeruginosa possesses a wide array of virulence traits, one of which is biofilm formation, which contributes to its pathogenicity. One of the main modulators of the P. aeruginosa-controlled intramembrane proteolysis pathway is AlgW, a membrane-bound periplasmic serine protease. In this work, we have used a set of density functional theory (DFT) calculations to understand the variety of chemical parameters in detail between andrographolide and levofloxacin, which show strong bactericidal activity against P. aeruginosa. Additionally, the stability and interaction of andrographolide and levofloxacin with the protein AlgW have been investigated by molecular docking and molecular dynamics (MD) simulations . Moreover, the growth and inhibition of biofilm production by P. aeruginosa experiments were also investigated, providing insight that andrographolide could be a potential natural product to inhibit P. aeruginosa.

Revealing the Interaction Mechanism between Mycobacterium tuberculosis GyrB and Novobiocin, SPR719 through Binding Thermodynamics and Dissociation Kinetics Analysis.

With the rapid emergence of drug-resistant strains of Mycobacterium tuberculosis (Mtb), various levels of resistance against existing anti-tuberculosis (TB) drugs have developed. Consequently, the identification of new anti-TB targets and drugs is critically urgent. DNA gyrase subunit B (GyrB) has been identified as a potential anti-TB target, with novobiocin and SPR719 proposed as inhibitors targeting GyrB. Therefore, elucidating the molecular interactions between GyrB and its inhibitors is crucial for the discovery and design of efficient GyrB inhibitors for combating multidrug-resistant TB. In this study, we revealed the detailed binding mechanisms and dissociation processes of the representative inhibitors, novobiocin and SPR719, with GyrB using classical molecular dynamics (MD) simulations, tau-random acceleration molecular dynamics (τ-RAMD) simulations, and steered molecular dynamics (SMD) simulations. Our simulation results demonstrate that both electrostatic and van der Waals interactions contribute favorably to the inhibitors' binding to GyrB, with Asn52, Asp79, Arg82, Lys108, Tyr114, and Arg141 being key residues for the inhibitors' attachment to GyrB. The τ-RAMD simulations indicate that the inhibitors primarily dissociate from the ATP channel. The SMD simulation results reveal that both inhibitors follow a similar dissociation mechanism, requiring the overcoming of hydrophobic interactions and hydrogen bonding interactions formed with the ATP active site. The binding and dissociation mechanisms of GyrB with inhibitors novobiocin and SPR719 obtained in our work will provide new insights for the development of promising GyrB inhibitors.

Evaluation of AlphaFold2 Structures for Hit Identification across Multiple Scenarios.

The introduction of AlphaFold2 (AF2) has sparked significant enthusiasm and generated extensive discussion within the scientific community, particularly among drug discovery researchers. Although previous studies have addressed the performance of AF2 structures in virtual screening (VS), a more comprehensive investigation is still necessary considering the paramount importance of structural accuracy in drug design. In this study, we evaluate the performance of AF2 structures in VS across three common drug discovery scenarios: targets with holo, apo, and AF2 structures; targets with only apo and AF2 structures; and targets exclusively with AF2 structures. We utilized both the traditional physics-based Glide and the deep-learning-based scoring function RTMscore to rank the compounds in the DUD-E, DEKOIS 2.0, and DECOY data sets. The results demonstrate that, overall, the performance of VS on AF2 structures is comparable to that on apo structures but notably inferior to that on holo structures across diverse scenarios. Moreover, when a target has solely AF2 structure, selecting the holo structure of the target from different subtypes within the same protein family produces comparable results with the AF2 structure for VS on the data set of the AF2 structures, and significantly better results than the AF2 structures on its own data set. This indicates that utilizing AF2 structures for docking-based VS may not yield most satisfactory outcomes, even when solely AF2 structures are available. Moreover, we rule out the possibility that the variations in VS performance between the binding pockets of AF2 and holo structures arise from the differences in their biological assembly composition.

Effects of Indocyanine Green (ICG) Imaging-Assisted Cholecystectomy on Intraoperative and Postoperative Complications: A meta-Analysis.

BACKGROUND: Accurate recognition of Calot's triangle during cholecystectomy is important in preventing intraoperative and postoperative complications. The use of indocyanine green (ICG) fluorescence imaging has become increasingly prevalent in cholecystectomy procedures. Our study aimed to evaluate the specific effects of ICG-assisted imaging in reducing complications. MATERIALS AND METHODS: A comprehensive search of databases including PubMed, Web of Science, Europe PMC, and WANFANGH DATA was conducted to identify relevant articles up to July 5, 2023. Review Manager 5.3 software was applied to statistical analysis. RESULTS: Our meta-analysis of 14 studies involving 3576 patients compared the ICG group (1351 patients) to the control group (2225 patients). The ICG group had a lower incidence of postoperative complications (4.78% vs 7.25%; RR .71; 95%CI: .54-.95; P = .02). Bile leakage was significantly reduced in the ICG group (.43% vs 2.02%; RR = .27; 95%CI: .12-.62; I2 = 0; P = .002), and they also had a lower bile duct drainage rate (24.8% vs 31.8% RR = .64, 95% CI: .44-.91, P = .01). Intraoperative complexes showed no statistically significant difference between the 2 groups (1.16% vs 9.24%; RR .17; 95%CI .03-1.02), but the incidence of intraoperative bleeding is lower in the ICG group. CONCLUSION: ICG fluorescence imaging-assisted cholecystectomy was associated with a range of benefits, including a lower incidence of postoperative complications, decreased rates of bile leakage, reduced bile duct drainage, fewer intraoperative complications, and reduced intraoperative bleeding.

TRIM5 as a promising diagnostic biomarker of hepatocellular carcinoma: integrated analysis and experimental validation.

The TRIM family is associated with the membrane, and its involvement in the progression, growth, and development of various cancer types has been researched extensively. However, the role played by the TRIM5 gene within this family has yet to be explored to a great extent in terms of hepatocellular carcinoma (HCC). The data of patients relating to mRNA expression and the survival rate of individuals diagnosed with HCC were extracted from The Cancer Genome Atlas (TCGA) database. UALCAN was employed to examine the potential link between TRIM5 expression and clinicopathological characteristics. In addition, enrichment analysis of differentially expressed genes (DEGs) was conducted as a means of deciphering the function and mechanism of TRIM5 in HCC. The data in the TCGA and TIMER2.0 databases was utilized to explore the correlation between TRIM5 and immune infiltration in HCC. WGCNA was performed as a means of assessing TRIM5-related co-expressed genes. The "OncoPredict" R package was also used for investigating the association between TRIM5 and drug sensitivity. Finally, qRT-PCR, Western blotting (WB) and immunohistochemistry (IHC) were employed for exploring the differential expression of TRIM5 and its clinical relevance in HCC. According to the results that were obtained from the vitro experiments, mRNA and protein levels of TRIM5 demonstrated a significant upregulation in HCC tissues. It is notable that TRIM5 expression levels were found to have a strong association with the infiltration of diverse immune cells and displayed a positive correlation with several immune checkpoint inhibitors. The TRIM5 expression also displayed promising clinical prognostic value for HCC patients.

Molecular mechanism of the one-component regulator RccR on bacterial metabolism and virulence.

The regulation of carbon metabolism and virulence is critical for the rapid adaptation of pathogenic bacteria to host conditions. In Pseudomonas aeruginosa, RccR is a transcriptional regulator of genes involved in primary carbon metabolism and is associated with bacterial resistance and virulence, although the exact mechanism is unclear. Our study demonstrates that PaRccR is a direct repressor of the transcriptional regulator genes mvaU and algU. Biochemical and structural analyses reveal that PaRccR can switch its DNA recognition mode through conformational changes triggered by KDPG binding or release. Mutagenesis and functional analysis underscore the significance of allosteric communication between the SIS domain and the DBD domain. Our findings suggest that, despite its overall structural similarity to other bacterial RpiR-type regulators, RccR displays a more complex regulatory element binding mode induced by ligands and a unique regulatory mechanism.

Enhancing Generic Reaction Yield Prediction through Reaction Condition-Based Contrastive Learning.

Deep learning (DL)-driven efficient synthesis planning may profoundly transform the paradigm for designing novel pharmaceuticals and materials. However, the progress of many DL-assisted synthesis planning (DASP) algorithms has suffered from the lack of reliable automated pathway evaluation tools. As a critical metric for evaluating chemical reactions, accurate prediction of reaction yields helps improve the practicality of DASP algorithms in the real-world scenarios. Currently, accurately predicting yields of interesting reactions still faces numerous challenges, mainly including the absence of high-quality generic reaction yield datasets and robust generic yield predictors. To compensate for the limitations of high-throughput yield datasets, we curated a generic reaction yield dataset containing 12 reaction categories and rich reaction condition information. Subsequently, by utilizing 2 pretraining tasks based on chemical reaction masked language modeling and contrastive learning, we proposed a powerful bidirectional encoder representations from transformers (BERT)-based reaction yield predictor named Egret. It achieved comparable or even superior performance to the best previous models on 4 benchmark datasets and established state-of-the-art performance on the newly curated dataset. We found that reaction-condition-based contrastive learning enhances the model's sensitivity to reaction conditions, and Egret is capable of capturing subtle differences between reactions involving identical reactants and products but different reaction conditions. Furthermore, we proposed a new scoring function that incorporated Egret into the evaluation of multistep synthesis routes. Test results showed that yield-incorporated scoring facilitated the prioritization of literature-supported high-yield reaction pathways for target molecules. In addition, through meta-learning strategy, we further improved the reliability of the model's prediction for reaction types with limited data and lower data quality. Our results suggest that Egret holds the potential to become an essential component of the next-generation DASP tools.

Experiment and molecular dynamics simulations reveal proanthocyanidin B2 and B3 can inhibit prion aggregation by different mechanisms.

Prion diseases are a group of fatal neurodegenerative diseases caused by the misfolding and aggregation of prion protein (PrP), and the inhibition of PrP aggregation is one of the most effective therapeutic strategies. Proanthocyanidin B2 (PB2) and B3 (PB3), the effective natural antioxidants have been evaluated for the inhibition of amyloid-related protein aggregation. Since PrP has similar aggregation mechanism with other amyloid-related proteins, will PB2 and PB3 affect the aggregation of PrP? In this paper, experimental and molecular dynamics (MD) simulation methods were combined to investigate the influence of PB2 and PB3 on PrP aggregation. Thioflavin T assays showed PB2 and PB3 could inhibit PrP aggregation in a concentrate-dependent manner in vitro. To understand the underlying mechanism, we performed 400 ns all-atom MD simulations. The results suggested PB2 could stabilize the α2 C-terminus and the hydrophobic core of protein by stabilizing two important salt bridges R156-E196 and R156-D202, and consequently made global structure of protein more stable. Surprisingly, PB3 could not stabilize PrP, which may inhibit PrP aggregation through a different mechanism. Since dimerization is the first step of aggregation, will PB3 inhibit PrP aggregation by inhibiting the dimerization? To verify our assumption, we then explored the effect of PB3 on protein dimerization by performing 800 ns MD simulations. The results suggested PB3 could reduce the residue contacts and hydrogen bonds between two monomers, preventing dimerization process of PrP. The possible inhibition mechanism of PB2 and PB3 on PrP aggregation could provide useful information for drug development against prion diseases.Communicated by Ramaswamy H. Sarma.

Causal associations between gut microbiota and primary biliary cholangitis: a bidirectional two-sample Mendelian randomization study.

Background: Previous studies have suggested an association between gut microbiota and primary biliary cholangitis (PBC). Nonetheless, the causal relationship between gut microbiota and PBC risk remains unclear. Methods: A bidirectional two-sample Mendelian Randomization (MR) study was employed using summary statistical data for gut microbiota and PBC from the MiBioGen consortium and Genome-Wide Association Studies (GWAS) database to investigate causal relationships between 211 gut microbiota and PBC risk. Inverse variance weighted (IVW) method was the primary analytical approach to assess causality, and the pleiotropy and heterogeneity tests were employed to verify the robustness of the findings. Additionally, we performed reverse MR analyses to investigate the possibility of the reverse causal association. Results: The IVW method identified five gut microbiota that demonstrated associations with the risk of PBC. Order Selenomonadales [odds ratio (OR) 2.13, 95% confidence interval (CI) 1.10-4.14, p = 0.03], Order Bifidobacteriales (OR 1.58, 95% CI 1.07-2.33, p = 0.02), and Genus Lachnospiraceae_UCG_004 (OR 1.64, 95%CI 1.06-2.55, p = 0.03) were correlated with a higher risk of PBC, while Family Peptostreptococcaceae (OR 0.65, 95%CI 0.43-0.98, p = 0.04) and Family Ruminococcaceae (OR 0.33, 95%CI 0.15-0.72, p = 0.01) had a protective effect on PBC. The reverse MR analysis demonstrated no statistically significant relationship between PBC and these five specific gut microbial taxa. Conclusion: This study revealed that there was a causal relationship between specific gut microbiota taxa and PBC, which may provide novel perspectives and a theoretical basis for the clinical prevention, diagnosis, and treatment of PBC.

Identification of Aggregation Mechanism of Acetylated PHF6* and PHF6 Tau Peptides Based on Molecular Dynamics Simulations and Markov State Modeling.

The microtubule-associated protein tau (MAPT) has a critical role in the development and preservation of the nervous system. However, tau's dysfunction and accumulation in the human brain can lead to several neurodegenerative diseases, such as Alzheimer's disease, Down's syndrome, and frontotemporal dementia. The microtubule binding (MTB) domain plays a significant, important role in determining the tau's pathophysiology, as the core of paired helical filaments PHF6* (275VQIINK280) and PHF6 (306VQIVYK311) of R2 and R3 repeat units, respectively, are formed in this region, which promotes tau aggregation. Post-translational modifications, and in particular lysine acetylation at K280 of PHF6* and K311 of PHF6, have been previously established to promote tau misfolding and aggregation. However, the exact aggregation mechanism is not known. In this study, we established an atomic-level nucleation-extension mechanism of the separated aggregation of acetylated PHF6* and PHF6 hexapeptides, respectively, of tau. We show that the acetylation of the lysine residues promotes the formation of ß-sheet enriched high-ordered oligomers. The Markov state model analysis of ac-PHF6* and ac-PHF6 aggregation revealed the formation of an antiparallel dimer nucleus which could be extended from both sides in a parallel manner to form mixed-oriented and high-ordered oligomers. Our study describes the detailed mechanism for acetylation-driven tau aggregation, which provides valuable insights into the effect of post-translation modification in altering the pathophysiology of tau hexapeptides.

CODD-Pred: A Web Server for Efficient Target Identification and Bioactivity Prediction of Small Molecules.

Target identification and bioactivity prediction are critical steps in the drug discovery process. Here we introduce CODD-Pred (COmprehensive Drug Design Predictor), an online web server with well-curated data sets from the GOSTAR database, which is designed with a dual purpose of predicting potential protein drug targets and computing bioactivity values of small molecules. We first designed a double molecular graph perception (DMGP) framework for target prediction based on a large library of 646 498 small molecules interacting with 640 human targets. The framework achieved a top-5 accuracy of over 80% for hitting at least one target on both external validation sets. Additionally, its performance on the external validation set comprising 200 molecules surpassed that of four existing target prediction servers. Second, we collected 56 targets closely related to the occurrence and development of cancer, metabolic diseases, and inflammatory immune diseases and developed a multi-model self-validation activity prediction (MSAP) framework that enables accurate bioactivity quantification predictions for small-molecule ligands of these 56 targets. CODD-Pred is a handy tool for rapid evaluation and optimization of small molecules with specific target activity. CODD-Pred is freely accessible at http://codd.iddd.group/.

Generic Interpretable Reaction Condition Predictions with Open Reaction Condition Datasets and Unsupervised Learning of Reaction Center.

Effective synthesis planning powered by deep learning (DL) can significantly accelerate the discovery of new drugs and materials. However, most DL-assisted synthesis planning methods offer either none or very limited capability to recommend suitable reaction conditions (RCs) for their reaction predictions. Currently, the prediction of RCs with a DL framework is hindered by several factors, including: (a) lack of a standardized dataset for benchmarking, (b) lack of a general prediction model with powerful representation, and (c) lack of interpretability. To address these issues, we first created 2 standardized RC datasets covering a broad range of reaction classes and then proposed a powerful and interpretable Transformer-based RC predictor named Parrot. Through careful design of the model architecture, pretraining method, and training strategy, Parrot improved the overall top-3 prediction accuracy on catalysis, solvents, and other reagents by as much as 13.44%, compared to the best previous model on a newly curated dataset. Additionally, the mean absolute error of the predicted temperatures was reduced by about 4 °C. Furthermore, Parrot manifests strong generalization capacity with superior cross-chemical-space prediction accuracy. Attention analysis indicates that Parrot effectively captures crucial chemical information and exhibits a high level of interpretability in the prediction of RCs. The proposed model Parrot exemplifies how modern neural network architecture when appropriately pretrained can be versatile in making reliable, generalizable, and interpretable recommendation for RCs even when the underlying training dataset may still be limited in diversity.

Artificial intelligence methods in kinase target profiling: Advances and challenges.

Kinases have a crucial role in regulating almost the full range of cellular processes, making them essential targets for therapeutic interventions against various diseases. Accurate kinase-profiling prediction is vital for addressing the selectivity/specificity challenges in kinase drug discovery, which is closely related to lead optimization, drug repurposing, and the understanding of potential drug side effects. In this review, we provide an overview of the latest advancements in machine learning (ML)-based and deep learning (DL)-based quantitative structure-activity relationship (QSAR) models for kinase profiling. We highlight current trends in this rapidly evolving field and discuss the existing challenges and future directions regarding experimental data set construction and model architecture design. Our aim is to offer practical insights and guidance for the development and utilization of these approaches.