DeepCPI:A Deep Learning-based Framework for Large-scale in silico Drug Screening

Accurate identification of compound–protein interactions(CPIs) in silico may deepen our understanding of the underlying mechanisms of drug action and thus remarkably facilitate drug discovery and development.Conventional similarity-or docking-based computational methods for predicting CPIs rarely exploit latent features from currently available large-scale unlabeled compound and protein data and often limit their usage to relatively small-scale datasets.In the present study,we propose Deep CPI,a novel general and scalable computational framework that combines effective feature embedding(a technique of representation learning) with powerful deep learning methods to accurately predict CPIs at a large scale.Deep CPI automatically learns the implicit yet expressive low-dimensional features of compounds and proteins from a massive amount of unlabeled data.Evaluations of the measured CPIs in large-scale databases,such as Ch EMBL and Binding DB,as well as of the known drug–target interactions from Drug Bank,demonstrated the superior predictive performance of Deep CPI.Furthermore,several interactions among smallmolecule compounds and three G protein-coupled receptor targets(glucagon-like peptide-1 receptor,glucagon receptor,and vasoactive intestinal peptide receptor) predicted using Deep CPI were experimentally validated.The present study suggests that Deep CPI is a useful and powerful tool for drug discovery and repositioning.The source code of Deep CPI can be downloaded from https://github.com/Fangping Wan/Deep CPI.     

Recent advances in target identification of bioactive natural products

Natural products are a tremendous source of tool discovery for basic science and drug discovery for clinical uses. In contrast to the large number of compounds isolated from nature, however, the number of compounds whose target molecules have been identified so far is fairly limited. Elucidation of the mechanism of how bioactive small molecules act in cells to induce biological activity (mode of action) is an attractive but challenging field of basic biology. At the same time, this is the major bottleneck for drug development of compounds identified in cell-based and phenotype-based screening. Although researchers’ experience and inspiration have been crucial for successful target identification, recent advancements in genomics, proteomics, and chemical genomics have made this challenging task possible in a systematic fashion.     

后基因组时代微生物天然产物高效发掘体系设计与展望

微生物天然产物具有丰富的化学结构多样性和诱人的生物活性,持续启迪着创新医药和农药的发现。近年来,随着高通量测序技术的快速发展,巨大的微生物基因组数据揭示了多样生物合成和新颖天然产物的潜能远高于以前的认知。然而,如何高效地激活隐性的生物合成基因簇 (BGCs) 并识别相应的化合物,以及避免已知代谢产物的重复发现等挑战依然严峻。本文描述了面对这些问题时基因组学、生物信息学、机器学习、代谢组学、基因编辑和合成生物学等新技术在发现药用先导化合物过程中提供的机遇;总结并论述了在潜力菌株优选、BGCs的生物信息学预测、沉默 BGCs的高效激活以及目标产物的识别和跟踪方面的新见解;提出了基于潜力菌株选择和多组学挖掘技术从微生物天然产物中高效发现先导结构的系统线路 (SPLSD),并讨论了未来天然产物药用先导发现的机遇和挑战。     

Biodiscovery from rare actinomycetes: an eco-taxonomical perspective

Microbial natural products, in particular, the ones produced by the members of the order Actinomycetales, will continue to represent an important route to the discovery of novel classes of bioactive compounds. As a result, the search for and discovery of lesser-known and/or novel actinomycetes is of significant interest to the industry due to a growing need for the development of new and potent therapeutic agents, mainly against drug resistant bacteria. Current advancements in genomics and metagenomics are adding strength to the target-directed search for detection and isolation of bioactive actinomycetes. New discoveries, however, will only stem from a sound understanding and interpretation of knowledge derived from conventional studies conducted since the discovery of streptomycin, on the ecology, taxonomy, physiology and metabolism of actinomycetes, and from a combination of this knowledge with currently available and continuously advancing molecular tools. Such a powerful information platform will then inevitably reveal the whereabouts, taxonomical and chemical identities of previously undetected bioactive actinomycetes including novel species of streptomycetes as potential producers of novel drug candidates.     

Target discovery in small-molecule cell-based screens by in situ proteome reactivity profiling

Chemical genomics aims to discover small molecules that affect biological processes through the perturbation of protein function. However, determining the protein targets of bioactive compounds remains a formidable challenge. We address this problem here through the creation of a natural product-inspired small-molecule library bearing protein-reactive elements. Cell-based screening identified a compound, MJE3, that inhibits breast cancer cell proliferation. In situ proteome reactivity profiling revealed that MJE3, but not other library members, covalently labeled the glycolytic enzyme phosphoglycerate mutase 1 (PGAM1), resulting in enzyme inhibition. Interestingly, MJE3 labeling and inhibition of PGAM1 were observed exclusively in intact cells. These results support the hypothesis that cancer cells depend on glycolysis for viability and promote PGAM1 as a potential therapeutic target. More generally, the incorporation of protein-reactive compounds into chemical genomics screens offers a means to discover targets of bioactive small molecules in living systems, thereby enabling downstream mechanistic investigations.     

Machine learning modeling of family wide enzyme-substrate specificity screens

Biocatalysis is a promising approach to sustainably synthesize pharmaceuticals, complex natural products, and commodity chemicals at scale. However, the adoption of biocatalysis is limited by our ability to select enzymes that will catalyze their natural chemical transformation on non-natural substrates. While machine learning and in silico directed evolution are well-posed for this predictive modeling challenge, efforts to date have primarily aimed to increase activity against a single known substrate, rather than to identify enzymes capable of acting on new substrates of interest. To address this need, we curate 6 different high-quality enzyme family screens from the literature that each measure multiple enzymes against multiple substrates. We compare machine learning-based compound-protein interaction (CPI) modeling approaches from the literature used for predicting drug-target interactions. Surprisingly, comparing these interaction-based models against collections of independent (single task) enzyme-only or substrate-only models reveals that current CPI approaches are incapable of learning interactions between compounds and proteins in the current family level data regime. We further validate this observation by demonstrating that our no-interaction baseline can outperform CPI-based models from the literature used to guide the discovery of kinase inhibitors. Given the high performance of non-interaction based models, we introduce a new structure-based strategy for pooling residue representations across a protein sequence. Altogether, this work motivates a principled path forward in order to build and evaluate meaningful predictive models for biocatalysis and other drug discovery applications.     

Identification of bioactive molecules by adipogenesis profiling of organic compounds

An important step in the postgenomic drug discovery is the construction of high quality chemical libraries that generate bioactive molecules at high rates. Here we report a cell-based approach to composing a focused library of biologically active compounds. A collection of bioactive non-cytotoxic chemicals was identified from a divergent library through the effects on the insulin-induced adipogenesis of 3T3-L1 cells, one of the most drastic and sensitive morphological alterations in cultured mammalian cells. The resulting focused library amply contained unique compounds with a broad range of pharmacological effects, including glucose-uptake enhancement, cytokine inhibition, osteogenesis stimulation, and selective suppression of cancer cells. Adipogenesis profiling of organic compounds generates a focused chemical library for multiple biological effects that are seemingly unrelated to adipogenesis, just as genetic screens with the morphology of fly eyes identify oncogenes and neurodegenerative genes.     

Tools and methodologies capable of isolating and identifying a target molecule for a bioactive compound

Elucidating the mechanism of action of bioactive compounds, such as commonly used pharmaceutical drugs and biologically active natural products, in the cells and the living body is important in drug discovery research. To this end, isolation and identification of target protein(s) for the bioactive compound are essential in understanding its function fully. And, development of reliable tools and methodologies capable of addressing efficiently identification and characterization of the target proteins based on the bioactive compounds accelerates drug discovery research. Affinity-based isolation and identification of target molecules for the bioactive compounds is a classic, but still powerful approach. This paper introduces recent progress on affinity chromatography system, focusing on development of practical affinity matrices and useful affinity-based methodologies on target identification. Beneficial affinity chromatography systems with using practical tools and useful methodologies facilitate chemical biology and drug discovery research.     

Structural genomics of GPCRs

G protein-coupled receptors (GPCRs) are targets for 60-70% of drugs in development today. Traditionally, the drug discovery process has relied on screening of chemical compounds to identify novel and more-efficient drug molecules. Structure-based drug design, however, provides a targeted approach but has been severely hampered by limited knowledge of high-resolution structures of GPCRs owing to the difficulties encountered in their expression, purification and crystallization. In addition to individual laboratories studying specific GPCRs, structural genomics initiatives have been established as large networks with a wide range of expertise in protein expression, purification and crystallography. Several of these national and international consortia have included GPCRs in their programs. Milligram quantities of GPCRs can now be expressed in several expression systems and purified to high homogeneity. However, success in crystallization still requires major technological improvement.     

Functional genomics: identifying drug targets for parasitic diseases

The genomic sequences of parasitic diseases are rapidly becoming available and, recently, the full sequence of Plasmodium falciparum has been published. Much has been promised from this genomic revolution including the identification of new drug targets and novel chemotherapeutic treatments for the control of parasitic diseases. The challenge to use this information efficiently will require functional genomics tools such as bioinformatics, microarrays, proteomics and chemical genomics to identify potential drug targets, and to allow the development of optimized lead compounds. The information generated from these tools will provide a crucial link from genomic analysis to drug discovery.     

Integrative genome‐scale metabolic analysis of Vibrio vulnificus for drug targeting and discovery

Although the genomes of many microbial pathogens have been studied to help identify effective drug targets and novel drugs, such efforts have not yet reached full fruition. In this study, we report a systems biological approach that efficiently utilizes genomic information for drug targeting and discovery, and apply this approach to the opportunistic pathogen Vibrio vulnificus CMCP6. First, we partially re‐sequenced and fully re‐annotated the V. vulnificus CMCP6 genome, and accordingly reconstructed its genome‐scale metabolic network, VvuMBEL943. The validated network model was employed to systematically predict drug targets using the concept of metabolite essentiality, along with additional filtering criteria. Target genes encoding enzymes that interact with the five essential metabolites finally selected were experimentally validated. These five essential metabolites are critical to the survival of the cell, and hence were used to guide the cost‐effective selection of chemical analogs, which were then screened for antimicrobial activity in a whole‐cell assay. This approach is expected to help fill the existing gap between genomics and drug discovery.     

Machine learning for target discovery in drug development

The discovery of macromolecular targets for bioactive agents is currently a bottleneck for the informed design of chemical probes and drug leads. Typically, activity profiling against genetically manipulated cell lines or chemical proteomics is pursued to shed light on their biology and deconvolute drug–target networks. By taking advantage of the ever-growing wealth of publicly available bioactivity data, learning algorithms now provide an attractive means to generate statistically motivated research hypotheses and thereby prioritize biochemical screens. Here, we highlight recent successes in machine intelligence for target identification and discuss challenges and opportunities for drug discovery.     

Proteomic methods for drug target discovery

The field of drug target discovery is currently very popular with a great potential for advancing biomedical research and chemical genomics. Innovative strategies have been developed to aid the process of target identification, either by elucidating the primary mechanism-of-action of a drug, by understanding side effects involving unanticipated 'off-target' interactions, or by finding new potential therapeutic value for an established drug. Several promising proteomic methods have been introduced for directly isolating and identifying the protein targets of interest that are bound by active small molecules or for visualizing enzyme activities affected by drug treatment. Significant progress has been made in this rapidly advancing field, speeding the clinical validation of drug candidates and the discovery of the novel targets for lead compounds developed using cell-based phenotypic screens. Using these proteomic methods, further insight into drug activity and toxicity can be ascertained.     

Chemical kinomics: a powerful strategy for target deconvolution

Kinomics is an emerging and promising approach for deciphering kinomes. Chemical kinomics is a discipline of chemical genomics that is also referred to as "chemogenomics", which is derived from chemistry and biology. Chemical kinomics has become a powerful approach to decipher complicated phosphorylation-based cellular signaling networks with the aid of small molecules that modulate kinase functions. Moreover, chemical kinomics has played a pivotal role in the field of kinase drug discovery as it enables identification of new molecular targets of small molecule kinase modulators and/or exploitation of novel functions of known kinases and has also provided novel chemical entities as hit/lead compounds. In this short review, contemporary chemical kinomics technologies such as activity-based protein profiling, T7 kinasetagged phages, kinobeads, three-hybrid systems, fluorescenttagged kinase binding assays, and chemical genomic profiling are discussed along with a novel allosteric Bcr-Abl kinase inhibitor (GNF-2/GNF-5) as a successful application of chemical kinomics approaches.     

Functional Knowledge Transfer for High-accuracy Prediction of Under-studied Biological Processes

A key challenge in genetics is identifying the functional roles of genes in pathways. Numerous functional genomics techniques (e.g. machine learning) that predict protein function have been developed to address this question. These methods generally build from existing annotations of genes to pathways and thus are often unable to identify additional genes participating in processes that are not already well studied. Many of these processes are well studied in some organism, but not necessarily in an investigator''s organism of interest. Sequence-based search methods (e.g. BLAST) have been used to transfer such annotation information between organisms. We demonstrate that functional genomics can complement traditional sequence similarity to improve the transfer of gene annotations between organisms. Our method transfers annotations only when functionally appropriate as determined by genomic data and can be used with any prediction algorithm to combine transferred gene function knowledge with organism-specific high-throughput data to enable accurate function prediction.We show that diverse state-of-art machine learning algorithms leveraging functional knowledge transfer (FKT) dramatically improve their accuracy in predicting gene-pathway membership, particularly for processes with little experimental knowledge in an organism. We also show that our method compares favorably to annotation transfer by sequence similarity. Next, we deploy FKT with state-of-the-art SVM classifier to predict novel genes to 11,000 biological processes across six diverse organisms and expand the coverage of accurate function predictions to processes that are often ignored because of a dearth of annotated genes in an organism. Finally, we perform in vivo experimental investigation in Danio rerio and confirm the regulatory role of our top predicted novel gene, wnt5b, in leftward cell migration during heart development. FKT is immediately applicable to many bioinformatics techniques and will help biologists systematically integrate prior knowledge from diverse systems to direct targeted experiments in their organism of study.     

Trends in application of advancing computational approaches in GPCR ligand discovery

G protein-coupled receptors (GPCRs) comprise the most important superfamily of protein targets in current ligand discovery and drug development. GPCRs are integral membrane proteins that play key roles in various cellular signaling processes. Therefore, GPCR signaling pathways are closely associated with numerous diseases, including cancer and several neurological, immunological, and hematological disorders. Computer-aided drug design (CADD) can expedite the process of GPCR drug discovery and potentially reduce the actual cost of research and development. Increasing knowledge of biological structures, as well as improvements on computer power and algorithms, have led to unprecedented use of CADD for the discovery of novel GPCR modulators. Similarly, machine learning approaches are now widely applied in various fields of drug target research. This review briefly summarizes the application of rising CADD methodologies, as well as novel machine learning techniques, in GPCR structural studies and bioligand discovery in the past few years. Recent novel computational strategies and feasible workflows are updated, and representative cases addressing challenging issues on olfactory receptors, biased agonism, and drug-induced cardiotoxic effects are highlighted to provide insights into future GPCR drug discovery.     

Chemical compound navigator: a web-based chem-BLAST, chemical taxonomy-based search engine for browsing compounds

A novel technique to annotate, query, and analyze chemical compounds has been developed and is illustrated by using the inhibitor data on HIV protease-inhibitor complexes. In this method, all chemical compounds are annotated in terms of standard chemical structural fragments. These standard fragments are defined by using criteria, such as chemical classification; structural, chemical, or functional groups; and commercial, scientific or common names or synonyms. These fragments are then organized into a data tree based on their chemical substructures. Search engines have been developed to use this data tree to enable query on inhibitors of HIV protease (http://xpdb.nist.gov/hivsdb/hivsdb.html). These search engines use a new novel technique, Chemical Block Layered Alignment of Substructure Technique (Chem-BLAST) to search on the fragments of an inhibitor to look for its chemical structural neighbors. This novel technique to annotate and query compounds lays the foundation for the use of the Semantic Web concept on chemical compounds to allow end users to group, sort, and search structural neighbors accurately and efficiently. During annotation, it enables the attachment of "meaning" (i.e., semantics) to data in a manner that far exceeds the current practice of associating "metadata" with data by creating a knowledge base (or ontology) associated with compounds. Intended users of the technique are the research community and pharmaceutical industry, for which it will provide a new tool to better identify novel chemical structural neighbors to aid drug discovery.     

Discovery of Novel Liver-Stage Antimalarials through Quantum Similarity

Without quantum theory any understanding of molecular interactions is incomplete. In principal, chemistry, and even biology, can be fully derived from non-relativistic quantum mechanics. In practice, conventional quantum chemical calculations are computationally too intensive and time consuming to be useful for drug discovery on more than a limited basis. A previously described, original, quantum-based computational process for drug discovery and design bridges this gap between theory and practice, and allows the application of quantum methods to large-scale in silico identification of active compounds. Here, we show the results of this quantum-similarity approach applied to the discovery of novel liver-stage antimalarials. Testing of only five of the model-predicted compounds in vitro and in vivo hepatic stage drug inhibition assays with P. berghei identified four novel chemical structures representing three separate quantum classes of liver-stage antimalarials. All four compounds inhibited liver-stage Plasmodium as a single oral dose in the quantitative PCR mouse liver-stage sporozoites-challenge model. One of the newly identified compounds, cethromycin [ABT-773], a macrolide-quinoline hybrid, is a drug with an extensive (over 5,000 people) safety profile warranting its exploitation as a new weapon for the current effort of malaria eradication. The results of our molecular modeling exceed current state-of-the-art computational methods. Drug discovery through quantum similarity is data-driven, agnostic to any particular target or disease process that can evaluate multiple phenotypic, target-specific, or co-crystal structural data. This allows the incorporation of additional pharmacological requirements, as well as rapid exploration of novel chemical spaces for therapeutic applications.     

Development of a high-content screening assay panel to accelerate mechanism of action studies for oncology research

Efficient elucidation of the biological mechanism of action of novel compounds remains a major bottleneck in the drug discovery process. To address this need in the area of oncology, we report the development of a multiparametric high-content screening assay panel at the level of single cells to dramatically accelerate understanding the mechanism of action of cell growth-inhibiting compounds on a large scale. Our approach is based on measuring 10 established end points associated with mitochondrial apoptosis, cell cycle disruption, DNA damage, and cellular morphological changes in the same experiment, across three multiparametric assays. The data from all of the measurements taken together are expected to help increase our current understanding of target protein functions, constrain the list of possible targets for compounds identified using phenotypic screens, and identify off-target effects. We have also developed novel data visualization and phenotypic classification approaches for detailed interpretation of individual compound effects and navigation of large collections of multiparametric cellular responses. We expect this general approach to be valuable for drug discovery across multiple therapeutic areas.     

Navigating large chemical spaces in early-phase drug discovery

The size of actionable chemical spaces is surging, owing to a variety of novel techniques, both computational and experimental. As a consequence, novel molecular matter is now at our fingertips that cannot and should not be neglected in early-phase drug discovery. Huge, combinatorial, make-on-demand chemical spaces with high probability of synthetic success rise exponentially in content, generative machine learning models go hand in hand with synthesis prediction, and DNA-encoded libraries offer new ways of hit structure discovery. These technologies enable to search for new chemical matter in a much broader and deeper manner with less effort and fewer financial resources.These transformational developments require new cheminformatics approaches to make huge chemical spaces searchable and analyzable with low resources, and with as little energy consumption as possible. Substantial progress has been made in the past years with respect to computation as well as organic synthesis. First examples of bioactive compounds resulting from the successful use of these novel technologies demonstrate their power to contribute to tomorrow's drug discovery programs. This article gives a compact overview of the state-of-the-art.