Showing papers by "Kaixian Chen published in 2020"
TL;DR: A new graph neural network architecture called Attentive FP for molecular representation that uses a graph attention mechanism to learn from relevant drug discovery datasets and achieves state-of-the-art predictive performances on a variety of datasets and that what it learns is interpretable.
Abstract: Hunting for chemicals with favorable pharmacological, toxicological, and pharmacokinetic properties remains a formidable challenge for drug discovery. Deep learning provides us with powerful tools to build predictive models that are appropriate for the rising amounts of data, but the gap between what these neural networks learn and what human beings can comprehend is growing. Moreover, this gap may induce distrust and restrict deep learning applications in practice. Here, we introduce a new graph neural network architecture called Attentive FP for molecular representation that uses a graph attention mechanism to learn from relevant drug discovery data sets. We demonstrate that Attentive FP achieves state-of-the-art predictive performances on a variety of data sets and that what it learns is interpretable. The feature visualization for Attentive FP suggests that it automatically learns nonlocal intramolecular interactions from specified tasks, which can help us gain chemical insights directly from data beyond human perception.
359 citations
TL;DR: New datasets specific for CPI prediction are constructed, a novel transformer neural network named TransformerCPI is proposed, and a more rigorous label reversal experiment is introduced to test whether a model learns true interaction features.
Abstract: Motivation Identifying compound-protein interaction (CPI) is a crucial task in drug discovery and chemogenomics studies, and proteins without three-dimensional structure account for a large part of potential biological targets, which requires developing methods using only protein sequence information to predict CPI. However, sequence-based CPI models may face some specific pitfalls, including using inappropriate datasets, hidden ligand bias and splitting datasets inappropriately, resulting in overestimation of their prediction performance. Results To address these issues, we here constructed new datasets specific for CPI prediction, proposed a novel transformer neural network named TransformerCPI, and introduced a more rigorous label reversal experiment to test whether a model learns true interaction features. TransformerCPI achieved much improved performance on the new experiments, and it can be deconvolved to highlight important interacting regions of protein sequences and compound atoms, which may contribute chemical biology studies with useful guidance for further ligand structural optimization. Availability and implementation https://github.com/lifanchen-simm/transformerCPI.
141 citations
TL;DR: 10i exhibited anti-neuroinflammatory effect in vitro and in vivo via inhibiting PKM2-mediated glycolysis and NLRP3 activation, indicating PKM 2 as a novel target for neuroinflammation and its related brain disorders.
Abstract: Benzoxepane derivatives were designed and synthesized, and one hit compound emerged as being effective in vitro with low toxicity. In vivo, this hit compound ameliorated both sickness behavior through anti-inflammation in LPS-induced neuroinflammatory mice model and cerebral ischemic injury through anti-neuroinflammation in rats subjected to transient middle cerebral artery occlusion. Target fishing for the hit compound using photoaffinity probes led to identification of PKM2 as the target protein responsible for anti-inflammatory effect of the hit compound. Furthermore, the hit exhibited an anti-neuroinflammatory effect in vitro and in vivo by inhibiting PKM2-mediated glycolysis and NLRP3 activation, indicating PKM2 as a novel target for neuroinflammation and its related brain disorders. This hit compound has a better safety profile compared to shikonin, a reported PKM2 inhibitor, identifying it as a lead compound in targeting PKM2 for the treatment of inflammation-related diseases.
62 citations
TL;DR: A virtual profiling model against a panel of 391 kinases based on large-scale bioactivity data and the multitask deep neural network algorithm is presented to create a comprehensive kinome interaction network for designing novel chemical modulators or drug repositioning.
Abstract: The kinome-wide virtual profiling of small molecules with high-dimensional structure-activity data is a challenging task in drug discovery. Here, we present a virtual profiling model against a panel of 391 kinases based on large-scale bioactivity data and the multitask deep neural network algorithm. The obtained model yields excellent internal prediction capability with an auROC of 0.90 and consistently outperforms conventional single-task models on external tests, especially for kinases with insufficient activity data. Moreover, more rigorous experimental validations including 1410 kinase-compound pairs showed a high-quality average auROC of 0.75 and confirmed many novel predicted "off-target" activities. Given the verified generalizability, the model was further applied to various scenarios for depicting the kinome-wide selectivity and the association with certain diseases. Overall, the computational model enables us to create a comprehensive kinome interaction network for designing novel chemical modulators or drug repositioning and is of practical value for exploring previously less studied kinases.
50 citations
TL;DR: An initially untargeted semisynthetic product, polycyclic lactone 11 with another novel carbocyclic skeleton was proved to be naturally occurring in the mollusk by comparative LC-MS/MS analysis, which provides a strategy to fish a vast number of novel trace components in Nature.
Abstract: Placobranchus ocellatus, regarded as a “solar-powered” sacoglossan mollusk, is well known to produce diverse and complex γ-pyrone polypropionates. Unexpectedly, in this study, the chemical investigation of P. ocellatus from the South China Sea led to the discovery of ocellatusones AD, a series of racemic non-γ-pyrone polyketides with novel skeletons, characterized by a bicyclo[3.2.1]octane (1, 2), a bicyclo[3.3.1]nonane (3) or a mesitylene-substituted dimethylfuran-3(2H)-one core (4). Extensive spectroscopic analysis, quantum chemical computation, chemical synthesis, and/or X-ray diffraction analysis were used to determine the structure and absolute configuration of the new compounds, including each enantiomer of racemic compounds 14 after chiral HPLC resolution. An important array of new and diversity-generating rearrangements is proposed to explain the biosynthesis of these unusual compounds based on careful structural analysis and comparison with six known co-occurring γ-pyrones (510). Furthermore, the successful biomimetic synthesis of ocellatusone A (1) from its γ-pyrone precursor (5) confirmed the biosynthetic relationship of these two compounds and the proposed rearrangement through an unprecedented acid promoted cascade reaction. Besides, an initially untargeted semisynthetic product, polycyclic lactone 11 with another novel carbocyclic skeleton was proved to be naturally occurring in the mollusk by comparative LC-MS/MS analysis, which provides a strategy to fish a vast number of novel trace components in Nature.
43 citations
TL;DR: In vivo pharmacodynamics results demonstrate that these synergetic effects caused by CaM‐PB NPs significantly contribute to the inhibition of tumor progression, demonstrating that theCaM‐ PB NPs with sequential theranostic functions are a promising system for effective cancer therapy.
29 citations
TL;DR: Four undescribed sulfur-containing indole alkaloids, isatisindigoticanines H, I and isatindigosides F, G along with three known analogues were obtained from Isatis tinctoria L. roots.
24 citations
TL;DR: In this article, a deep learning model was trained to learn the relationships between the chemical contexts, reaction conditions and product yields based on high-quality existing experimental data, and then extrapolate reasonably to unseen reactions by in silico exploration of accessible reaction space.
Abstract: Here we report a feasibility study of a deep learning model for exploring the optimal reaction conditions for given chemical reactions. The model was trained to learn the relationships between the chemical contexts, reaction conditions and product yields based on high-quality existing experimental data, and then extrapolate reasonably to unseen reactions by in silico exploration of accessible reaction space. This strategy was applied to the Suzuki–Miyaura cross-coupling reaction to find the best catalysts for given reactants and at the same time to discover the optimum combination of the reaction conditions. We demonstrated that the trained model was able to determine the productive catalysts as well as the most favorable catalyst loading and reaction temperature for both modeled reactions and external unseen reactions. This work aims to provide an insight into the feasibility of introducing a deep learning method in the optimization of chemical reaction conditions.
18 citations
TL;DR: α-Diazo quinones were applied in an Ir(iii)-catalyzed direct C-H functionalization assisted by N-phenylacetamide for the construction of highly functionalized 2-hydroxy-2'-amino-1,1'-biaryl scaffolds in good to excellent yields.
17 citations
TL;DR: The synergistic effect of the heat shock protein 90 inhibitor 17-AAG and the histone deacetylase 6 inhibitor Belinostat in triple-negative breast cancer MDA-MB-231 cells is reported, by detection of proliferation, apoptosis and cell cycle arrest following treatment with this combination.
Abstract: Breast cancer is one of the most common malignancies that threaten the health of women. Although there are a few chemotherapies for the clinical treatment of breast cancer, these therapies are faced with the problems of drug‑resistance and metastasis. Drug combination can help to reduce the adverse side effects of chemotherapies using single drugs, and also help to overcome common drug‑resistance during clinical treatment of breast cancer. The present study reported the synergistic effect of the heat shock protein 90 inhibitor 17‑AAG and the histone deacetylase 6 inhibitor Belinostat in triple‑negative breast cancer (TNBC) MDA‑MB‑231 cells, by detection of proliferation, apoptosis and cell cycle arrest following treatment with this combination. Subsequently, RNA sequencing (RNA‑seq) data was collected and analyzed to investigate the synergistic mechanism of this combination. Based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) signaling pathways revealed by RNA‑seq data analysis, a wound‑healing assay was used to investigate the effect of this combination on the migration of MDA‑MB‑231 cells. Compared with treatment with 17‑AAG or Belinostat alone, both the viability inhibition and apoptosis rate of MDA‑MB‑231 cells were significantly enhanced in the combination group. The combination index values were <1 in three concentration groups. Revealed by the RNA‑seq data analysis, the most significantly enriched KEGG pathways in the combination group were closely associated with cell migration. Based on these findings, the anti‑migration effect of this combination was investigated. It was revealed that the migration of MDA‑MB‑231 cells was significantly suppressed in the combination group compared with in the groups treated with 17‑AAG or Belinostat alone. In terms of specific genes, the mRNA expression levels of TEA domain family proteins were significantly decreased in the combination group, whereas the phosphorylation of YY1 associated protein 1 and modulator of VRAC current 1 was significantly enhanced in the combination group. These alterations may help to explain the anti‑migration effect of this combination. Belinostat has already been approved as a treatment for T‑cell lymphoma and 17‑AAG is undergoing clinical trials. These findings could provide a beneficial reference for the clinical treatment of patients with TNBC.
17 citations
TL;DR: A palladium-catalyzed [4 + 3] dearomatizing cycloaddition of N-iminoquinolinium ylides has been developed for the construction of saturated seven-membered ring in good yields and features mild reaction conditions and good functional group tolerance.
Abstract: In this study, a palladium-catalyzed [4 + 3] dearomatizing cycloaddition of N-iminoquinolinium ylides has been developed for the construction of saturated seven-membered ring in good yields. This approach features mild reaction conditions and good functional group tolerance. Furthermore, gram-scale preparation and transformation were carried out, which further underscored its synthetic utility and applicability.
TL;DR: A machine learning algorithm has been developed that predicts the conversion rate for the DNA-compatible reaction of a building block with a model DNA-conjugate, allowing for a challenging reaction, with an otherwise very low building block pass rate in the test reaction, to still be used in DEL synthesis.
TL;DR: Eight bisindole alkaloids including six undescribed ones (1a/1b-5) were isolated from an alcohol extract of the Isatis indigotica roots and were determined to be a pair of enantiomers with a ratio of approximately 1:1 by chiral-phase chromatography.
TL;DR: A traceless approach to quinolin-4(1H)-one scaffolds through Rh(III)-catalyzed redox-neutral [3+3] cyclization of N-nitrosoanilines with cyclopropenones has been achieved, which features short reaction time and atom-economical combination without extra additives.
Abstract: A traceless approach to quinolin-4(1H)-one scaffolds through Rh(III)-catalyzed redox-neutral [3+3] cyclization of N-nitrosoanilines with cyclopropenones has been achieved. This protocol features short reaction time and atom-economical combination without extra additives, which can be further applied in the construction of privileged heterocyclic compounds in pharmaceutical chemistry.
TL;DR: DC-CPin711 showed potent in vitro inhibitory activities to CBP bromodomain with a decent selectivity towards BRD4 bromODomains and good cellular activity to leukemia cells, which could further be applied to related biological and translational studies as well as serve as a lead compound for future development of potent and selective CBPs.
TL;DR: USP8 inhibitors with novel scaffold with broad prospects for being a probe for USP8-related academic and clinical research are discovered via high throughput screening based on Ubiquitin-rhodamine-110 (Ubiquit in-Rho-110) fluorometric activity assay.
TL;DR: DC-CPin734 showing good potency, selectivity and anti AML activity could serve as a potent and selective in vitro and in vivo probe of CBP bromodomain and a promising lead compound for future drug development.
TL;DR: In this article, three new indole alkaloid glycosides were obtained from the roots of Isatis indigotica and their putative biosynthetic pathways were proposed.
TL;DR: Linariifolioside II and (2S)‐2‐hydroxy‐5‐oxoproline methyl ester (2), two new compounds along with 13 known compounds were obtained from the aerial part of Pseudolysimachion linariIFolium Holub subsp.
Abstract: Linariifolioside II (1) and (2S)-2-hydroxy-5-oxoproline methyl ester (2), two new compounds along with 13 known compounds were obtained from the aerial part of Pseudolysimachion linariifolium Holub subsp. dilatatum (Nakai & Kitag.) D.Y. Hong. Their chemical structures were revealed mainly through NMR and MS data. The absolute configuration of 2 was deduced by comparing its experimental CD with the calculated ECD spectra. At a concentration of 1 mm, total antioxidant capacities of compounds 1-15 were measured using a rapid ABTS method in vitro. Compounds 1, 3-5, and 11-14 exhibited approximately equal antioxidant capacity to that of vitamin C (Vc).
TL;DR: An improved computational model integrating both atom-level and molecule-level features is developed to predict whether a drug-like molecule is a potential human AOX (hAOX) substrate and to identify the corresponding sites of metabolism.
Abstract: Aldehyde oxidase (AOX) is a drug metabolizing molybdo-flavoenzyme that has gained increasing attention because of contribution to the biotransformation in phase I metabolism of xenobiotics. Unfortunately, the intra- and interspecies variations in AOX activity and lack of reliable and predictive animal models make evaluation of AOX-catalyzed metabolism prone to be misleading. In this study, we developed an improved computational model integrating both atom-level and molecule-level features to predict whether a drug-like molecule is a potential human AOX (hAOX) substrate and to identify the corresponding sites of metabolism. Additionally, we combined the proposed computational strategy and in vitro experiments for evaluating the metabolic property of a series of epigenetic-related drug candidates still in the early stage of development. In summary, this study provides an improved strategy to evaluate the liability of molecules toward hAOX and offers useful information for accelerating the drug design and optimization stage.
TL;DR: The discovery of triazoloquinoxaline 1a is described as a novel STING agonist possessing the potential to be further developed for antiviral and antitumor treatment via Structure-based Virtual Screening.
TL;DR: Excessive reactive oxygen species is highly involved in inflammation and the pathogenesis of infectious diseases as a potent scavenger of ROS, peroxiredo.
Abstract: Excessive reactive oxygen species (ROS) is highly involved in inflammation and the pathogenesis of infectious diseases As a potent scavenger of ROS, peroxiredo
TL;DR: A novel Siamese spectral-based graph convolutional network model for inferring the protein targets of chemical compounds from gene transcriptional profiles that was successfully trained to learn from known compound-target pairs by uncovering the hidden correlations between compound perturbation profiles and gene knockdown profiles.
Abstract: Computational target fishing aims to investigate the mechanism of action or the side effects of bioactive small molecules. Unfortunately, conventional ligand-based computational methods only explore a confined chemical space, and structure-based methods are limited by the availability of crystal structures. Moreover, these methods cannot describe cellular context-dependent effects and are thus not useful for exploring the targets of drugs in specific cells. To address these challenges, we propose a novel Siamese spectral-based graph convolutional network (SSGCN) model for inferring the protein targets of chemical compounds from gene transcriptional profiles. Although the gene signature of a compound perturbation only provides indirect clues of the interacting targets, the SSGCN model was successfully trained to learn from known compound-target pairs by uncovering the hidden correlations between compound perturbation profiles and gene knockdown profiles. Using a benchmark set, the model achieved impressive target inference results compared with previous methods such as Connectivity Map and ProTINA. More importantly, the powerful generalization ability of the model observed with the external LINCS phase II dataset suggests that the model is an efficient target fishing or repositioning tool for bioactive compounds.
TL;DR: Computational and experimental results revealed the intrinsic conformational change of cIAP1, which could not only explain previously identified key mutations, but also be exploited for further design and development of anti-tumor compounds that target the cI AP1 protein.
Abstract: Inhibitor of apoptosis proteins (IAPs) are important regulators of apoptosis, and protein targets for the development of anti-cancer drugs. Cellular inhibitor of apoptosis protein-1 (cIAP1) is an i...
TL;DR: In this paper, chemical fractionation of the ethanolic extract of the roots of Isatis tinctoria L (I tinctorsia) yielded fourteen indole alkaloids including four new ones, isatisindigoticanine L-N and isatindigo
TL;DR: In this article, the authors summarize the characteristics of publicly accessible genomic databases and discuss the trends of artificial intelligence applications in drug sensitivity prediction for cancer cell lines, including machine learning, networks and multimodal deep neural networks.
Abstract: The development of computational methods for the prediction of effective therapeutic strategies based on the genomic information of patients is the main challenge of precision medicine. Since the 21st century, next-generation sequencing (NGS) has opened up new possibilities for personalized medicine. Extensive characterization at the molecular level for hundreds of cancer cell lines has been brought to the public eye by many organizations and agencies around the world. For example, the National Cancer Institute 60 Human Cancer Cell Line Screen (NCI-60), Cancer Cell Line Encyclopedia (CCLE) and Genomics of Drug Sensitivity in Cancer (GDSC) have provided large-scale omics data such as genomic, transcriptomic and epigenomic data characterizing cancer cell lines, and The Cancer Genome Atlas (TCGA) has molecularly characterized over 20000 primary cancers of patients. Combined with the drug response data of cancer cell lines, multiomics data could be used to analyse the mechanisms of action of anticancer drugs, which could be incorporated into precision medicine strategies. Over several decades, artificial intelligence (AI) technologies based on big data have revolutionized bioinformatics. AI has built a bridge between genomics and drug sensitivity by promoting the development of predictive models for the drug response of cancer cell lines. The 2012 NCI-DREAM drug prediction challenge has been particularly influential, as the innovative applications of machine learning that emerged from it have laid the groundwork for future studies. However, classic machine learning models are still challenging in terms of predictability because they limit the systematic integration of high-dimensional multiomics data. Therefore, network-based approaches, including link prediction and network representation, have become mainstream methods for drug response prediction. On the one hand, network-based approaches have not faced the “small n, large p” problem since the multiomics features are either represented in a gene/protein network or embedded in similarity networks between cell lines. On the other hand, the introduction of gene regulatory networks (GRNs) and protein-protein interactions (PPIs) into the predictive model can provide a functional background for the integration of genomic data and thereby improve the predictive performance of drug response. In addition to network-based approaches, multimodal deep learning models can systematically integrate multiomic data by considering them as different modalities. Generally, there are three feature fusion methods in deep neural networks: Input-level feature fusion (early fusion), intermediate feature fusion and decision-level fusion (late fusion). Intermediate feature fusion is predominant in drug response prediction studies, by which features are learned separately for each type of omics data and then integrated into one unified representation to be used as the input for a classifier or a regressor. Moreover, the features of drug structures can be used as a model to improve the performance. In brief, we summarize the characteristics of publicly accessible genomic databases and discuss the trends of artificial intelligence applications in drug sensitivity prediction for cancer cell lines, including machine learning, networks and multimodal deep neural networks.
TL;DR: An accurate prediction model for bioactivity was built with consensus method, which was superior to all individual models and should be a valuable tool for lead optimization.
Abstract: Background Enhancing a compound's biological activity is the central task for lead optimization in small molecules drug discovery. However, it is laborious to perform many iterative rounds of compound synthesis and bioactivity tests. To address the issue, it is highly demanding to develop high quality in silico bioactivity prediction approaches, to prioritize such more active compound derivatives and reduce the trial-and-error process. Methods Two kinds of bioactivity prediction models based on a large-scale structure-activity relationship (SAR) database were constructed. The first one is based on the similarity of substituents and realized by matched molecular pair analysis, including SA, SA_BR, SR, and SR_BR. The second one is based on SAR transferability and realized by matched molecular series analysis, including Single MMS pair, Full MMS series, and Multi single MMS pairs. Moreover, we also defined the application domain of models by using the distance-based threshold. Results Among seven individual models, Multi single MMS pairs bioactivity prediction model showed the best performance (R2 = 0.828, MAE = 0.406, RMSE = 0.591), and the baseline model (SA) produced the most lower prediction accuracy (R2 = 0.798, MAE = 0.446, RMSE = 0.637). The predictive accuracy could further be improved by consensus modeling (R2 = 0.842, MAE = 0.397 and RMSE = 0.563). Conclusion An accurate prediction model for bioactivity was built with a consensus method, which was superior to all individual models. Our model should be a valuable tool for lead optimization.
Patent•
21 Sep 2020
TL;DR: In this paper, the polymorph of maleate (R)-methyl (2-(3-aminopiperidin-1-yl)-3-(2-cyanobenzyl)-4-carbonyl-3,4-dihydrothiophene [3,2-d] pyrimidin 6-carboxylic acid and methods for preparing it and a pharmaceutical composition were presented.
Abstract: FIELD: medicine; pharmaceuticals.SUBSTANCE: invention relates to the polymorph of maleate (R)-methyl-2-(3-aminopiperidin-1-yl)-3-(2-cyanobenzyl)-4-carbonyl-3,4-dihydrothiophene [3,2-d] pyrimidin-6-carboxylic acid and methods for preparing it and a pharmaceutical composition. Polymorph is a crystal having high stability and low hygroscopicity, wherein the crystalline form is selected from crystalline form A, crystalline form B and crystalline form C.EFFECT: crystalline form has hypoglycemic activity and can be used to produce a new drug for treating or preventing type II diabetes mellitus and/or complications of type II diabetes mellitus.10 cl, 7 ex, 18 dwg