Models of mild cognitive deficits in risk assessment in early psychosis.

BACKGROUND: Mild cognitive deficits (MCD) emerge before the first episode of psychosis (FEP) and persist in the clinical high-risk (CHR) stage. This study aims to refine risk prediction by developing MCD models optimized for specific early psychosis stages and target populations. METHODS: A comprehensive neuropsychological battery assessed 1059 individuals with FEP, 794 CHR, and 774 matched healthy controls (HCs). CHR subjects, followed up for 2 years, were categorized into converters (CHR-C) and non-converters (CHR-NC). The MATRICS Consensus Cognitive Battery standardized neurocognitive tests were employed. RESULTS: Both the CHR and FEP groups exhibited significantly poorer performance compared to the HC group across all neurocognitive tests (all p < 0.001). The CHR-C group demonstrated poorer performance compared to the CHR-NC group on three sub-tests: visuospatial memory (p < 0.001), mazes (p = 0.005), and symbol coding (p = 0.023) tests. Upon adjusting for sex and age, the performance of the MCD model was excellent in differentiating FEP from HC, as evidenced by an Area Under the Receiver Operating Characteristic Curve (AUC) of 0.895 (p < 0.001). However, when applied in the CHR group for predicting CHR-C (AUC = 0.581, p = 0.008), the performance was not satisfactory. To optimize the efficiency of psychotic risk assessment, three distinct MCD models were developed to distinguish FEP from HC, predict CHR-C from CHR-NC, and identify CHR from HC, achieving accuracies of 89.3%, 65.6%, and 80.2%, respectively. CONCLUSIONS: The MCD exhibits variations in domains, patterns, and weights across different stages of early psychosis and diverse target populations. Emphasizing precise risk assessment, our findings highlight the importance of tailored MCD models for different stages and risk levels.

Constructing Innovative Covalent and Noncovalent Compound Libraries: Insights from 3D Protein-Ligand Interactions.

Noncovalent interactions between small-molecule drugs and protein targets assume a pivotal role in drug design. Moreover, the design of covalent inhibitors, forming covalent bonds with amino acid residues, requires rational reactivity for their covalent warheads, presenting a key challenge as well. Understanding the intricacies of these interactions provides a more comprehensive understanding of molecular binding mechanisms, thereby guiding the rational design of potent inhibitors. In this study, we adopted the fragment-based drug design approach, introducing a novel methodology to extract noncovalent and covalent fragments according to distinct three-dimensional (3D) interaction modes from noncovalent and covalent compound libraries. Additionally, we systematically replaced existing ligands with rational fragment substitutions, based on the spatial orientation of fragments in 3D space. Furthermore, we adopted a molecular generation approach to create innovative covalent inhibitors. This process resulted in the recombination of a noncovalent compound library and several covalent compound libraries, constructed by two commonly encountered covalent amino acids: cysteine and serine. We utilized noncovalent ligands in KLIFS and covalent ligands in CovBinderInPDB as examples to recombine noncovalent and covalent libraries. These recombined compound libraries cover a substantial portion of the chemical space present in the original compound libraries and exhibit superior performance in terms of molecular scaffold diversity compared to the original compound libraries and other 11 commercial libraries. We also recombined BTK-focused libraries, and 23 compounds within our libraries have been validated by former researchers to possess potential biological activity. The establishment of these compound libraries provides valuable resources for virtual screening of covalent and noncovalent drugs targeting similar molecular targets.

Correlative relationship between body mass index and heart rate variability in psychiatric disorders.

OBJECTIVE: Indicators of heart rate variability (HRV) have been used to assess the autonomic activity. However, the influence of obesity on HRV in these patients remains to be determined. This study aimed to examine how obesity (measured with the body mass index [BMI]) affects HRV and determine whether the effect varies among different psychiatric disorders. We recruited 3159 consecutive patients, including 1744 with schizophrenia, 966 with mood disorders, and 449 with anxiety disorders. Patients were divided into four groups based on BMI: underweight (< 18.5), normal weight (18.5-23.9), overweight (24-27.9), and obese (≥ 28). The cardiovascular status was assessed using several time- and frequency-based HRV indicators, measured via electrocardiogram signals recorded for 5 min. The mean BMI of the participants was 23.6 ± 4.0. The patients in the overweight and obese groups were 29.4% and 13.6% of the total, respectively. The HRV indicators were higher in underweight and normal-weight patients than in the overweight and obese ones. After stratification based on the psychiatric diagnosis, the patients with mood disorders showed lower HRV than those with schizophrenia or anxiety disorder in the normal-weight group. In contrast, in the overweight and obese groups the patients with mood disorders showed higher HRV than those with the other disorders. The HRV variables were significantly associated with BMI, and higher BMI was associated with higher heart rates and lower HRV. These results indicate that weight gain in psychiatric disorders is associated with an imbalance in autonomic nerve activity. However, the relationship between autonomic activity, weight gain, and psychiatric disorders warrants further investigation.

Comparison of Direct Pars Repair Techniques for Spondylolysis in Young Patients: Pedicle Screw Hook System versus Pedicle Screw Rod.

Background: Lumbar spondylolysis (LS) poses a potential threat, and there is a need to evaluate and compare the effectiveness of direct pars repair techniques. Objective: To assess and compare the clinical and radiographic outcomes of direct pars repair techniques using the pedicle screw hook system (PSHS) and the pedicle screw rod system (PSRS) in young symptomatic patients with lumbar spondylolysis. Methods: A retrospective study was conducted to compare clinical and radiological data in young symptomatic LS patients after surgery. Records of 45 post-surgery LS patients with a minimum 24-month follow-up (January 2014 to June 2019) were reviewed. A total of 26 patients underwent PSHS, and 19 had PSRS. Treatment outcomes were analyzed using the visual analog pain scale (VAS), Oswestry disability index (ODI), MacNab criteria, lumbar fusion status, and Pfirrmann grading standards. Patient baseline characteristics were also compared between the two groups. Results: No disc degeneration was observed in either PSHS or PSRS groups at 24 months postoperatively, according to the Pfirrmann grading scale. The PSRS group outperformed the PSHS group in operative time, intraoperative blood loss, postoperative drainage, length of hospital stays, ODI, VAS values at 3 months postoperatively, and fusion status at 6 months postoperatively. No notable differences were observed in other parameters during the 24-month follow-up period, and no significant surgical complications were recorded. Conclusions: Direct pars repair techniques using PSHS and PSRS yielded satisfactory clinical and radiographic results in young patients with symptomatic LS. PSRS, compared to PSHS, demonstrated greater effectiveness in young individuals with LS and promoted early recovery.

Association between serum cytokines and timeframe for conversion from clinical high-risk to psychosis.

AIM: Although many studies have explored the link between inflammatory markers and psychosis, there is a paucity of research investigating the temporal progression in individuals at clinical high-risk (CHR) who eventually develop full psychosis. To address this gap, we investigated the correlation between serum cytokine levels and Timeframe for Conversion to Psychosis (TCP) in individuals with CHR. METHODS: We enrolled 53 individuals with CHR who completed a 5-year follow-up with a confirmed conversion to psychosis. Granulocyte macrophage-colony stimulating factor (GM-CSF), interleukin (IL)-1ß, 2, 6, 8, 10, tumor necrosis factor-α (TNF-α), and vascular endothelial growth factor (VEGF) levels were measured at baseline and 1-year. Correlation and quantile regression analyses were performed. RESULTS: The median TCP duration was 14 months. A significantly shorter TCP was associated with higher levels of TNF-α (P = 0.022) and VEGF (P = 0.016). A negative correlation was observed between TCP and TNF-α level (P = 0.006) and VEGF level (P = 0.04). Quantile regression indicated negative associations between TCP and GM-CSF levels below the 0.5 quantile, IL-10 levels below the 0.3 quantile, IL-2 levels below the 0.25 quantile, IL-6 levels between the 0.65 and 0.75 quantiles, TNF-α levels below the 0.8 quantile, and VEGF levels below the 0.7 quantile. A mixed linear effects model identified significant time effects for IL-10 and IL-2, and significant group effects for changes in IL-2 and TNF-α. CONCLUSIONS: Our findings underscore that a more pronounced baseline inflammatory state is associated with faster progression of psychosis in individuals with CHR. This highlights the importance of considering individual inflammatory profiles during early intervention and of tailoring preventive measures for risk profiles.

Suppression of Cation Intermixing Highly Boosts the Performance of Core-Shell Lanthanide Upconversion Nanoparticles.

Lanthanide upconversion nanoparticles (UCNPs) have been extensively explored as biomarkers, energy transducers, and information carriers in wide-ranging applications in areas from healthcare and energy to information technology. In promoting the brightness and enriching the functionalities of UCNPs, core-shell structural engineering has been well-established as an important approach. Despite its importance, a strong limiting issue has been identified, namely, cation intermixing in the interfacial region of the synthesized core-shell nanoparticles. Currently, there still exists confusion regarding this destructive phenomenon and there is a lack of facile means to reach a delicate control of it. By means of a new set of experiments, we identify and provide in this work a comprehensive picture for the major physical mechanism of cation intermixing occurring in synthesis of core-shell UCNPs, i.e., partial or substantial core nanoparticle dissolution followed by epitaxial growth of the outer layer and ripening of the entire particle. Based on this picture, we provide an easy but effective approach to tackle this issue that enables us to produce UCNPs with highly boosted optical properties.

Serum angioneurin levels following electroconvulsive therapy for mood disorders.

OBJECTIVES: The efficacy of electroconvulsive therapy (ECT) in treating mood disorders (MDs) is hypothesized to be mediated by the induction of neurotrophic factors (denoted "angioneurins") that trigger neuronal plasticity. This study aimed to assess the effects of ECT on serum angioneurin levels in patients with MD. METHODS: A total of 110 patients with MDs including 30 with unipolar depression, 25 with bipolar depression (BD), 55 with bipolar mania (BM), and 50 healthy controls were included in the study. Patients were subdivided into two groups: those who received ECT + medication (12 ECT sessions) and those who received only medication (no-ECT). Depressive and manic symptom assessments and measurements of vascular endothelial growth factor (VEGF), fibroblast growth factor-2, nerve growth factor (NGF), and insulin-like growth factor-1 levels in blood samples were performed at baseline and week 8. RESULTS: Patients in the ECT group, specifically those with BD and BM, had significantly increased levels of VEGF compared to their baseline VEGF levels (p = 0.002). No significant changes in angioneurin levels were observed in the no-ECT group. Serum NGF levels were significantly associated with a reduction in depressive symptoms. Angioneurin levels were not associated with manic symptom reduction. CONCLUSIONS: This study hints that ECT may increase VEGF levels with angiogenic mechanisms that amplify NGF signaling to promote neurogenesis. It may also contribute to changes in brain function and emotional regulation. However, further animal experiments and clinical validation are needed.

Changes in Inflammatory Markers in Clinical High Risk of Developing Psychosis.

INTRODUCTION: Immune alterations are associated with the progression of psychosis. However, there are few studies designed to longitudinally measure inflammatory biomarkers during psychotic episodes. We aimed to assess changes in biomarkers from the prodromal phase to psychotic episodes in individuals with clinical high risk (CHR) of psychosis and compare converters and non-converters to psychosis as well as healthy controls (HCs). METHODS: We enrolled 394 individuals with CHR and 100 HCs. A total of 263 individuals with CHR completed the 1-year follow-up, and 47 had converted to psychosis. Interleukin (IL)-1ß, 2, 6, 8, 10, tumor necrosis factor-α (TNF-α), and vascular endothelial growth factor levels were measured at baseline and 1 year after completion of the clinical assessment. RESULTS: The baseline serum levels of IL-10, IL-2, and IL-6 were significantly lower in the conversion group than in the non-conversion group (IL-10, p = 0.010; IL-2, p = 0.023; IL-6, p = 0.012) and HC (IL-6: p = 0.034). Self-controlled comparisons showed that IL-2 changed significantly (p = 0.028), and IL-6 levels tended toward significance (p = 0.088) in the conversion group. In the non-conversion group, serum levels of TNF-α (p = 0.017) and VEGF (p = 0.037) changed significantly. Repeated measures analysis of variance revealed a significant time effect related to TNF-α (F = 4.502, p = 0.037, effect size (η2) = 0.051), a group effect related to IL-1ß (F = 4.590, p = 0.036, η2 = 0.062), and IL-2 (F = 7.521, p = 0.011, η2 = 0.212), but no time × group effect. DISCUSSION: Alterations in the serum levels of inflammatory cytokines were found to precede the first episode of psychosis in the CHR population, particularly for those who later converted to psychosis. Longitudinal analysis supports the varied roles of cytokines in individuals with CHR with later psychotic conversion or non-conversion outcomes.

An Interpretable Multitask Framework BiLAT Enables Accurate Prediction of Cyclin-Dependent Protein Kinase Inhibitors.

The cyclin-dependent protein kinases (CDKs) are protein-serine/threonine kinases with crucial effects on the regulation of cell cycle and transcription. CDKs can be a hallmark of cancer since their excessive expression could lead to impaired cell proliferation. However, the selectivity profile of most developed CDK inhibitors is not enough, which have hindered the therapeutic use of CDK inhibitors. In this study, we propose a multitask deep learning framework called BiLAT based on SMILES representation for the prediction of the inhibitory activity of molecules on eight CDK subtypes (CDK1, 2, 4-9). The framework is mainly composed of an improved bidirectional long short-term memory module BiLSTM and the encode layer of the Transformer framework. Additionally, the data enhancement method of SMILES enumeration is applied to improve the performance of the model. Compared with baseline predictive models based on three conventional machine learning methods and two multitask deep learning algorithms, BiLAT achieves the best performance with the highest average AUC, ACC, F1-score, and MCC values of 0.938, 0.894, 0.911, and 0.715 for the test set. Moreover, we constructed a targeted external data set CDK-Dec for the CDK family, which mainly contains bait values screened by 3D similarity with active compounds. This dataset was utilized in the subsequent evaluation of our model. It is worth mentioning that the BiLAT model is interpretable and can be used by chemists to design and synthesize compounds with improved activity. To further verify the generalization ability of the multitask BiLAT model, we also conducted another evaluation on three public datasets (Tox21, ClinTox, and SIDER). Compared with several currently popular models, BiLAT shows the best performance on two datasets. These results indicate that BiLAT is an effective tool for accelerating drug discovery.

RPBP: Deep Retrosynthesis Reaction Prediction Based on Byproducts.

Retrosynthesis prediction is crucial in organic synthesis and drug discovery, aiding chemists in designing efficient synthetic routes for target molecules. Data-driven deep retrosynthesis prediction has gained importance due to new algorithms and enhanced computing power. Although existing models show certain predictive power on the USPTO-50K benchmark data set, no one considers the effects of byproducts during the prediction process, which may be due to the lack of byproduct information in the benchmark data set. Here, we propose a novel two-stage retrosynthesis reaction prediction framework based on byproducts called RPBP. First, RPBP predicts the byproduct involved in the reaction based on the product molecule. Then, it handles an end-to-end prediction problem based on the prediction of reactants by product and byproduct. Unlike other methods that first identify the potential reaction center and then predict reactant molecules, RPBP considers additional information from byproducts, such as reaction reagents, conditions, and sites. Interestingly, adding byproducts reduces model learning complexity in natural language processing (NLP). Our RPBP model achieves 54.7% and 66.6% top-1 retrosynthesis prediction accuracy when the reaction class is unknown and known, respectively. It outperforms existing methods for known-class reactions, thanks to the rich chemical information in byproducts. The prediction of four kinase drugs from the literature demonstrates the model's practicality and potential to accelerate drug discovery.

Attenuated niacin response is associated with a subtype of first-episode drug-naïve psychosis characterized as serious negative symptoms.

Although the phenomenon of attenuated niacin response (ANR) has been widely replicated in some patients with first-episode psychosis (FEP), its relevance to the negative symptoms (NS) of psychosis remains unclear. Total of 240 patients with drug-naïve FEP and 101 healthy controls (HCs) were recruited, and 209 were followed up for 1 year. Psychotic symptoms were assessed using the Positive and Negative Syndrome Scale (PANSS), and niacin-induced responses were measured using laser Doppler flowmetry. We calculated the log-transform EC50 [concentration of methyl nicotinate required to elicit a half-maximal blood flow (MBF) response] and MBF values. Core-NS was generated by factor analysis of the PANSS-NS subscale and cluster analysis to produce subtypes. Significant differences were found in the log10 (EC50) values between the FEP and HC groups (p < 0.001), supporting the ANR in patients with FEP. A higher NS severity was found in the ANR subgroup than that in other patients. Factor analysis determined that a two-dimensional model included core NS and rigidity of thinking. The log10 (EC50) value was significantly associated with only the core NS. Cluster analysis revealed three subtypes-36.7% (cluster-1, n = 88), 16.7% (cluster-2, n = 40), and 46.7% (cluster-3, n = 112). Cluster-2 characterized by extensive NS appeared to have a more remarkable ANR and less symptomatic improvement than those with other clusters during follow-up. No significant changes were found in the niacin response trajectories between the baseline and follow-up. Our findings indicate a significant correlation between ANR and core NS in patients with FEP. ANR may be a potential biomarker for certain subtypes with NS-dominated characteristics and poor symptomatic remission.

Eye movement indices as predictors of conversion to psychosis in individuals at clinical high risk.

Eye movement abnormalities have been established as an "endophenotype" of schizophrenia. However, less is known about the possibility of these abnormalities as biomarkers for psychosis conversion among clinical high risk (CHR) populations. In the present study, 108 CHR individuals and 70 healthy controls (HC) underwent clinical assessments and eye-tracking tests, comprising fixation stability and free-viewing tasks. According to three-year follow-up outcomes, CHR participants were further stratified into CHR-converter (CHR-C; n = 21) and CHR-nonconverter (CHR-NC; n = 87) subgroups. Prediction models were constructed using Cox regression and logistic regression. The CHR-C group showed more saccades of the fixation stability test (no distractor) and a reduced saccade amplitude of the free-viewing test than HC. Moreover, the CHR-NC group exhibited excessive saccades and an increased saccade amplitude of the fixation stability test (no distractor; with distractor) compared with HC. Furthermore, two indices could effectively discriminate CHR-C from CHR-NC with an area under the receiver-operating characteristic (ROC) curve of 0.80, including the saccade number of the fixation stability test (no distractor) and the saccade amplitude of the free-viewing test. Combined with negative symptom scores of the Scale of Prodromal Symptoms, the area was 0.81. These findings support that eye movement alterations might emerge before the onset of clinically overt psychosis and could assist in predicting psychosis transition among CHR populations.

Multi-channel GCN ensembled machine learning model for molecular aqueous solubility prediction on a clean dataset.

This study constructed a new aqueous solubility dataset and a solubility regression model which was ensembled by GCN and machine learning models. Aqueous solubility is a key physiochemical property of small molecules in drug discovery. In the past few decades, there have been many studies about solubility prediction. However, many of these studies have high root mean squared error (RMSE). Meanwhile, their dataset always contains salt compounds and solubility data obtained from different experimental conditions. In this paper, we constructed a clean dataset with 2609 compounds, which was small but contains only solubility records without salts at the same temperatures (25 °C). Here, we applied graph convolutional neural network (GCN) to construct an aqueous solubility prediction model. To enhance the performance of the model, the molecular MACCS key fingerprints and physiochemical descriptors were also combined with the GCN model to build a multi-channel model. Additionally, the authors also built two machine learning models (support vector regression and gradient boost decision tree) and assembled them to the GCN model to improve the root mean squared error (RMSE = 0.665). Finally, comparative experiments have shown that our framework achieved the best performance on ESOL dataset (RMSEval = 0.56, RMSEtest = 0.44) and surpassed four established software on aqueous solubility prediction of new compounds.

Multimodal multi-task deep neural network framework for kinase-target prediction.

Kinase plays a significant role in various disease signaling pathways. Due to the highly conserved sequence of kinase family members, understanding the selectivity profile of kinase inhibitors remains a priority for drug discovery. Previous methods for kinase selectivity identification use biochemical assays, which are very useful but limited by the protein available. The lack of kinase selectivity can exert benefits but also can cause adverse effects. With the explosion of the dataset for kinase activities, current computational methods can achieve accuracy for large-scale selectivity predictions. Here, we present a multimodal multi-task deep neural network model for kinase selectivity prediction by calculating the fingerprint and physiochemical descriptors. With the multimodal inputs of structure and physiochemical properties information, the multi-task framework could accurately predict the kinome map for selectivity analysis. The proposed model displays better performance for kinase-target prediction based on system evaluations.

Combined Fluorescence Fluctuation and Spectrofluorometric Measurements Reveal a Red-Shifted, Near-IR Emissive Photo-Isomerized Form of Cyanine 5.

Cyanine fluorophores are extensively used in fluorescence spectroscopy and imaging. Upon continuous excitation, especially at excitation conditions used in single-molecule and super-resolution experiments, photo-isomerized states of cyanines easily reach population probabilities of around 50%. Still, effects of photo-isomerization are largely ignored in such experiments. Here, we studied the photo-isomerization of the pentamethine cyanine 5 (Cy5) by two similar, yet complementary means to follow fluorophore blinking dynamics: fluorescence correlation spectroscopy (FCS) and transient-state (TRAST) excitation-modulation spectroscopy. Additionally, we combined TRAST and spectrofluorimetry (spectral-TRAST), whereby the emission spectra of Cy5 were recorded upon different rectangular pulse-train excitations. We also developed a framework for analyzing transitions between multiple emissive states in FCS and TRAST experiments, how the brightness of the different states is weighted, and what initial conditions that apply. Our FCS, TRAST, and spectral-TRAST experiments showed significant differences in dark-state relaxation amplitudes for different spectral detection ranges, which we attribute to an additional red-shifted, emissive photo-isomerized state of Cy5, not previously considered in FCS and single-molecule experiments. The photo-isomerization kinetics of this state indicate that it is formed under moderate excitation conditions, and its population and emission may thus deserve also more general consideration in fluorescence imaging and spectroscopy experiments.

Automatic auditory processing features in distinct subtypes of patients at clinical high risk for psychosis: Forecasting remission with mismatch negativity.

Individuals at clinical high risk (CHR) for psychosis exhibit a compromised mismatch negativity (MMN) response, which indicates dysfunction of pre-attentive deviance processing. Event-related potential and time-frequency (TF) information, in combination with clinical and cognitive profiles, may provide insight into the pathophysiology and psychopathology of the CHR stage and predict the prognosis of CHR individuals. A total of 92 individuals with CHR were recruited and followed up regularly for up to 3 years. Individuals with CHR were classified into three clinical subtypes demonstrated previously, specifically 28 from Cluster 1 (characterized by extensive negative symptoms and cognitive deficits), 31 from Cluster 2 (characterized by thought and behavioral disorganization, with moderate cognitive impairment), and 33 from Cluster 3 (characterized by the mildest symptoms and cognitive deficits). Auditory MMN to frequency and duration deviants was assessed. The event-related spectral perturbation (ERSP) and inter-trial coherence (ITC) were acquired using TF analysis. Predictive indices for remission were identified using logistic regression analyses. As expected, reduced frequency MMN (fMMN) and duration MMN (dMMN) responses were noted in Cluster 1 relative to the other two clusters. In the TF analysis, Cluster 1 showed decreased theta and alpha ITC in response to deviant stimuli. The regression analyses revealed that dMMN latency and alpha ERSP to duration deviants, theta ITC to frequency deviants and alpha ERSP to frequency deviants, and fMMN latency were significant MMN predictors of remission for the three clusters. MMN variables outperformed behavioral variables in predicting remission of Clusters 1 and 2. Our findings indicate relatively disrupted automatic auditory processing in a certain CHR subtype and a close affinity between these electrophysiological indexes and clinical profiles within different clusters. Furthermore, MMN indexes may serve as predictors of subsequent remission from the CHR state. These findings suggest that the auditory MMN response is a potential neurophysiological marker for distinct clinical subtypes of CHR.

Effective Reaction-Based De Novo Strategy for Kinase Targets: A Case Study on MERTK Inhibitors.

Reaction-based de novo design is the computational generation of novel molecular structures by linking building blocks using reaction vectors derived from chemistry knowledge. In this work, we first adopted a recurrent neural network (RNN) model to generate three groups of building blocks with different functional groups and then constructed an in silico target-focused combinatorial library based on chemical reaction rules. Mer tyrosine kinase (MERTK) was used as a study case. Combined with a scaffold enrichment analysis, 15 novel MERTK inhibitors covering four scaffolds were achieved. Among them, compound 5a obtained an IC50 value of 53.4 nM against MERTK without any further optimization. The efficiency of hit identification could be significantly improved by shrinking the compound library with the fragment iterative optimization strategy and enriching the dominant scaffold in the hinge region. We hope that this strategy can provide new insights for accelerating the drug discovery process.

KID: A Kinase-Focused Interaction Database and Its Application in the Construction of Kinase-Focused Molecule Databases.

Protein kinases are important drug targets for the treatment of several diseases. The interaction between kinases and ligands is vital in the process of small-molecule kinase inhibitor (SMKI) design. In this study, we propose a method to extract fragments and amino acid residues from crystal structures for kinase-ligand interactions. In addition, core fragments that interact with the important hinge region of kinases were extracted along with their decorations. Based on the superimposed structural data of kinases from the kinase-ligand interaction fingerprint and structure database, we obtained two libraries, namely, a hinge-unfocused fragment-amino acid pair library (FAP Lib) that contains 6672 pairs of fragments and corresponding amino-acids, and a hinge-focused hinge binder library (HB Lib) of 3560 pairs of hinge-binding scaffolds with their corresponding decorations. These two libraries constitute a kinase-focused interaction database (KID). In depth analysis was conducted on KID to explore important characteristics of fragments in the design of SMKIs. With KID, we built two kinase-focused molecule databases, one called Recomb_DB, which contains 1,72,346 molecules generated through fragment recombination based on the FAP Lib, and another called RsdHB_DB, which contains 93,030 molecules generated based on our HB Lib using molecular generation methods. Compared with five databases both commercial and non-commercial, these two databases both ranked top 3 in scaffold diversity, top 4 in molecule fingerprint diversity, and are more focused on the chemical space of kinase inhibitors. Hence, KID presents a useful addition to existing databases for the exploration of novel SMKIs.

Accurate calculation of absolute free energy of binding for SHP2 allosteric inhibitors using free energy perturbation.

Accurate prediction of binding affinity is a primary objective in structure-based drug discovery. A free energy perturbation (FEP) method based on molecular dynamics simulation shows great promise for protein-ligand binding affinity predictions. However, accurate calculation of binding affinity for allosteric inhibitors remains unknown and elusive, which hampers the discovery of allosteric inhibitors. Allosteric inhibitors exhibit several significant advantages over orthosteric inhibitors including higher specificity and lower side effects. Allosteric inhibitors against SHP2 are thought to be beneficial not only for diseases related to metabolism, but also for cancer, which make SHP2 a potential drug target. However, high structural sensitivity makes structural optimization of SHP2 allosteric inhibitors face challenges. Herein, we calculated the absolute binding free energy of SHP2 allosteric inhibitors using the FEP method by employing different λ-windows/simulation time sampling strategies. A simulation run with 32 λ-windows/64 ps sampling strategy delivered an excellent correlation (r = 0.96) and an unprecedented low mean absolute error of 0.5 kcal mol-1 between predicted binding free energies and experimental ones, outperforming the MM/PBSA method. Our study demonstrates the possibility to accurately calculate the absolute binding free energy of allosteric inhibitors using FEP, which offers exciting prospects for the discovery of more effective allosteric inhibitors.

Discovery of Dual FGFR4 and EGFR Inhibitors by Machine Learning and Biological Evaluation.

Kinase inhibitors are widely used in antitumor research, but there are still many problems such as drug resistance and off-target toxicity. A more suitable solution is to design a multitarget inhibitor with certain selectivity. Herein, computational and experimental studies were applied to the discovery of dual inhibitors against FGFR4 and EGFR. A quantitative structure-property relationship (QSPR) study was carried out to predict the FGFR4 and EGFR activity of a data set consisting of 843 and 5088 compounds, respectively. Four different machine learning methods including support vector machine (SVM), random forest (RF), gradient boost regression tree (GBRT), and XGBoost (XGB) were built using the most suitable features selected by the mutual information algorithm. As for FGFR4 and EGFR, SVM showed the best performance with R2test-FGFR4 = 0.80 and R2test-EGFR = 0.75, demonstrating excellent model stability, which was used to predict the activity of some compounds from an in-house database. Finally, compound 1 was selected, which exhibits inhibitory activity against FGFR4 (IC50 = 86.2 nM) and EGFR (IC50 = 83.9 nM) kinase, respectively. Furthermore, molecular docking and molecular dynamics simulations were performed to identify key amino acids for the interaction of compound 1 with FGFR4 and EGFR. In this paper, the machine-learning-based QSAR models were established and effectively applied to the discovery of dual-target inhibitors against FGFR4 and EGFR, demonstrating the great potential of machine learning strategies in dual inhibitor discovery.