DeepCOP: deep learning-based approach to predict gene regulating effects of small molecules.

MOTIVATION: Recent advances in the areas of bioinformatics and chemogenomics are poised to accelerate the discovery of small molecule regulators of cell development. Combining large genomics and molecular data sources with powerful deep learning techniques has the potential to revolutionize predictive biology. In this study, we present Deep gene COmpound Profiler (DeepCOP), a deep learning based model that can predict gene regulating effects of low-molecular weight compounds. This model can be used for direct identification of a drug candidate causing a desired gene expression response, without utilizing any information on its interactions with protein target(s). RESULTS: In this study, we successfully combined molecular fingerprint descriptors and gene descriptors (derived from gene ontology terms) to train deep neural networks that predict differential gene regulation endpoints collected in LINCS database. We achieved 10-fold cross-validation RAUC scores of and above 0.80, as well as enrichment factors of >5. We validated our models using an external RNA-Seq dataset generated in-house that described the effect of three potent antiandrogens (with different modes of action) on gene expression in LNCaP prostate cancer cell line. The results of this pilot study demonstrate that deep learning models can effectively synergize molecular and genomic descriptors and can be used to screen for novel drug candidates with the desired effect on gene expression. We anticipate that such models can find a broad use in developing novel cancer therapeutics and can facilitate precision oncology efforts. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

ETS transcription factors as emerging drug targets in cancer.

The ETS family of proteins consists of 28 transcription factors, many of which have been implicated in development and progression of a variety of cancers. While one family member, ERG, has been rigorously studied in the context of prostate cancer where it plays a critical role, other ETS factors keep emerging as potential hallmark oncodrivers. In recent years, numerous studies have reported initial discoveries of small molecule inhibitors of ETS proteins and opened novel avenues for ETS-directed cancer therapies. This review summarizes the state of the art data on therapeutic targeting of ETS family members and highlights the corresponding drug discovery strategies.

Computer-Aided Discovery of Small Molecule Inhibitors of Thymocyte Selection-Associated High Mobility Group Box Protein (TOX) as Potential Therapeutics for Cutaneous T-Cell Lymphomas.

Cutaneous T-cell lymphomas (CTCL) are the most common primary lymphomas of the skin. We have previously identified thymocyte selection-associated high mobility group (HMG) box protein (TOX) as a promising drug target in CTCL; however, there are currently no small molecules able to directly inhibit TOX. We aimed to address this unmet opportunity by developing anti-TOX therapeutics with the use of computer-aided drug discovery methods. The available NMR-resolved structure of the TOX protein was used to model its DNA-binding HMG-box domain. To investigate the druggability of the corresponding protein-DNA interface on TOX, we performed a pilot virtual screening of 200,000 small molecules using in silico docking and identified 'hot spots' for drug-binding on the HMG-box domain. We then performed a large-scale virtual screening of 7.6 million drug-like compounds that were available from the ZINC15 database. As a result, a total of 140 top candidate compounds were selected for subsequent in vitro validation. Of those, 18 small molecules have been characterized as selective TOX inhibitors.

Toxic Colors: The Use of Deep Learning for Predicting Toxicity of Compounds Merely from Their Graphic Images.

The majority of computational methods for predicting toxicity of chemicals are typically based on "nonmechanistic" cheminformatics solutions, relying on an arsenal of QSAR descriptors, often vaguely associated with chemical structures, and typically employing "black-box" mathematical algorithms. Nonetheless, such machine learning models, while having lower generalization capacity and interpretability, typically achieve a very high accuracy in predicting various toxicity endpoints, as unambiguously reflected by the results of the recent Tox21 competition. In the current study, we capitalize on the power of modern AI to predict Tox21 benchmark data using merely simple 2D drawings of chemicals, without employing any chemical descriptors. In particular, we have processed rather trivial 2D sketches of molecules with a supervised 2D convolutional neural network (2DConvNet) and demonstrated that the modern image recognition technology results in prediction accuracies comparable to the state-of-the-art cheminformatics tools. Furthermore, the performance of the image-based 2DConvNet model was comparatively evaluated on an external set of compounds from the Prestwick chemical library and resulted in experimental identification of significant and previously unreported antiandrogen potentials for several well-established generic drugs.

Mitf is a master regulator of the v-ATPase, forming a control module for cellular homeostasis with v-ATPase and TORC1.

The v-ATPase is a fundamental eukaryotic enzyme that is central to cellular homeostasis. Although its impact on key metabolic regulators such as TORC1 is well documented, our knowledge of mechanisms that regulate v-ATPase activity is limited. Here, we report that the Drosophila transcription factor Mitf is a master regulator of this holoenzyme. Mitf directly controls transcription of all 15 v-ATPase components through M-box cis-sites and this coordinated regulation affects holoenzyme activity in vivo. In addition, through the v-ATPase, Mitf promotes the activity of TORC1, which in turn negatively regulates Mitf. We provide evidence that Mitf, v-ATPase and TORC1 form a negative regulatory loop that maintains each of these important metabolic regulators in relative balance. Interestingly, direct regulation of v-ATPase genes by human MITF also occurs in cells of the melanocytic lineage, showing mechanistic conservation in the regulation of the v-ATPase by MITF family proteins in fly and mammals. Collectively, this evidence points to an ancient module comprising Mitf, v-ATPase and TORC1 that serves as a dynamic modulator of metabolism for cellular homeostasis.

Selectively targeting the DNA-binding domain of the androgen receptor as a prospective therapy for prostate cancer.

The androgen receptor (AR) is a transcription factor that has a pivotal role in the occurrence and progression of prostate cancer. The AR is activated by androgens that bind to its ligand-binding domain (LBD), causing the transcription factor to enter the nucleus and interact with genes via its conserved DNA-binding domain (DBD). Treatment for prostate cancer involves reducing androgen production or using anti-androgen drugs to block the interaction of hormones with the AR-LBD. Eventually the disease changes into a castration-resistant form of PCa where LBD mutations render anti-androgens ineffective or where constitutively active AR splice variants, lacking the LBD, become overexpressed. Recently, we identified a surfaced exposed pocket on the AR-DBD as an alternative drug-target site for AR inhibition. Here, we demonstrate that small molecules designed to selectively bind the pocket effectively block transcriptional activity of full-length and splice variant AR forms at low to sub-micromolar concentrations. The inhibition is lost when residues involved in drug interactions are mutated. Furthermore, the compounds did not impede nuclear localization of the AR and blocked interactions with chromatin, indicating the interference of DNA binding with the nuclear form of the transcription factor. Finally, we demonstrate the inhibition of gene expression and tumor volume in mouse xenografts. Our results indicate that the AR-DBD has a surface site that can be targeted to inhibit all forms of the AR, including enzalutamide-resistant and constitutively active splice variants and thus may serve as a potential avenue for the treatment of recurrent and metastatic prostate cancer.

The mycobacterial antibiotic resistance determinant WhiB7 acts as a transcriptional activator by binding the primary sigma factor SigA (RpoV).

Tuberculosis therapeutic options are limited by the high intrinsic antibiotic resistance of Mycobacterium tuberculosis. The putative transcriptional regulator WhiB7 is crucial for the activation of systems that provide resistance to diverse antibiotic classes. Here, we used in vitro run-off, two-hybrid assays, as well as mutagenic, complementation and protein pull-down experiments, to characterize WhiB7 as an auto-regulatory, redox-sensitive transcriptional activator in Mycobacterium smegmatis. We provide the first direct biochemical proof that a WhiB protein promotes transcription and also demonstrate that this activity is sensitive to oxidation (diamide). Its partner protein for transcriptional activation was identified as SigA, the primary sigma factor subunit of RNA polymerase. Residues required for the interaction mapped to region 4 of SigA (including R515H) or adjacent domains of WhiB7 (including E63D). WhiB7's ability to provide a specific spectrum of antibiotic-resistance was dependent on these residues as well as its C-terminal AT-hook module that binds to an AT-rich motif immediately upstream of the -35 hexamer recognized by SigA. These experimentally established constrains, combined with protein structure predictions, were used to generate a working model of the WhiB7-SigA-promoter complex. Inhibitors preventing WhiB7 interactions could allow the use of previously ineffective antibiotics for treatment of mycobacterial diseases.

Prioritizing disease-linked variants, genes, and pathways with an interactive whole-genome analysis pipeline.

Whole-genome sequencing (WGS) studies are uncovering disease-associated variants in both rare and nonrare diseases. Utilizing the next-generation sequencing for WGS requires a series of computational methods for alignment, variant detection, and annotation, and the accuracy and reproducibility of annotation results are essential for clinical implementation. However, annotating WGS with up to date genomic information is still challenging for biomedical researchers. Here, we present one of the fastest and highly scalable annotation, filtering, and analysis pipeline-gNOME-to prioritize phenotype-associated variants while minimizing false-positive findings. Intuitive graphical user interface of gNOME facilitates the selection of phenotype-associated variants, and the result summaries are provided at variant, gene, and genome levels. Moreover, the enrichment results of specific variants, genes, and gene sets between two groups or compared with population scale WGS datasets that is already integrated in the pipeline can help the interpretation. We found a small number of discordant results between annotation software tools in part due to different reporting strategies for the variants with complex impacts. Using two published whole-exome datasets of uveal melanoma and bladder cancer, we demonstrated gNOME's accuracy of variant annotation and the enrichment of loss-of-function variants in known cancer pathways. gNOME Web server and source codes are freely available to the academic community (http://gnome.tchlab.org).

Polycomb repressive complex 2 regulates normal development of the mouse heart.

RATIONALE: Epigenetic marks are crucial for organogenesis, but their role in heart development is poorly understood. Polycomb repressive complex 2 (PRC2) trimethylates histone H3 at lysine 27, which establishes H3K27me3 repressive epigenetic marks that promote tissue-specific differentiation by silencing ectopic gene programs. OBJECTIVE: We studied the function of PRC2 in murine heart development using a tissue-restricted conditional inactivation strategy. METHODS AND RESULTS: Inactivation of the PRC2 subunit Ezh2 by Nkx2-5(Cre) (Ezh2(NK)) caused lethal congenital heart malformations, namely, compact myocardial hypoplasia, hypertrabeculation, and ventricular septal defect. Candidate and genome-wide RNA expression profiling and chromatin immunoprecipitation analyses of Ezh2(NK) heart identified genes directly repressed by EZH2. Among these were the potent cell cycle inhibitors Ink4a/b (inhibitors of cyclin-dependent kinase 4 A and B), the upregulation of which was associated with decreased cardiomyocyte proliferation in Ezh2(NK). EZH2-repressed genes were enriched for transcriptional regulators of noncardiomyocyte expression programs such as Pax6, Isl1, and Six1. EZH2 was also required for proper spatiotemporal regulation of cardiac gene expression, because Hcn4, Mlc2a, and Bmp10 were inappropriately upregulated in ventricular RNA. PRC2 was also required later in heart development, as indicated by cardiomyocyte-restricted TNT-Cre inactivation of the PRC2 subunit Eed. However, Ezh2 inactivation by TNT-Cre did not cause an overt phenotype, likely because of functional redundancy with Ezh1. Thus, early Ezh2 inactivation by Nk2-5(Cre) caused later disruption of cardiomyocyte gene expression and heart development. CONCLUSIONS: Our study reveals a previously undescribed role of EZH2 in regulating heart formation and shows that perturbation of the epigenetic landscape early in cardiogenesis has sustained disruptive effects at later developmental stages.

Clinically-observed FOXA1 mutations upregulate SEMA3C through transcriptional derepression in prostate cancer.

FOXA1 is a pioneer transcription factor that is frequently mutated in prostate, breast, bladder, and salivary gland malignancies. Indeed, metastatic castration-resistant prostate cancer (mCRPC) commonly harbour FOXA1 mutations with a prevalence of 35%. However, despite the frequent recurrence of FOXA1 mutations in prostate cancer, the mechanisms by which FOXA1 variants drive its oncogenic effects are still unclear. Semaphorin 3C (SEMA3C) is a secreted autocrine growth factor that drives growth and treatment resistance of prostate and other cancers and is known to be regulated by both AR and FOXA1. In the present study, we characterize FOXA1 alterations with respect to its regulation of SEMA3C. Our findings reveal that FOXA1 alterations lead to elevated levels of SEMA3C both in prostate cancer specimens and in vitro. We further show that FOXA1 negatively regulates SEMA3C via intronic cis elements, and that mutations in FOXA1 forkhead domain attenuate its inhibitory function in reporter assays, presumably by disrupting DNA binding of FOXA1. Our findings underscore the key role of FOXA1 in prostate cancer progression and treatment resistance by regulating SEMA3C expression and suggest that SEMA3C may be a driver of growth and tumor vulnerability of mCRPC harboring FOXA1 alterations.

gSearch: a fast and flexible general search tool for whole-genome sequencing.

BACKGROUND: Various processes such as annotation and filtering of variants or comparison of variants in different genomes are required in whole-genome or exome analysis pipelines. However, processing different databases and searching among millions of genomic loci is not trivial. RESULTS: gSearch compares sequence variants in the Genome Variation Format (GVF) or Variant Call Format (VCF) with a pre-compiled annotation or with variants in other genomes. Its search algorithms are subsequently optimized and implemented in a multi-threaded manner. The proposed method is not a stand-alone annotation tool with its own reference databases. Rather, it is a search utility that readily accepts public or user-prepared reference files in various formats including GVF, Generic Feature Format version 3 (GFF3), Gene Transfer Format (GTF), VCF and Browser Extensible Data (BED) format. Compared to existing tools such as ANNOVAR, gSearch runs more than 10 times faster. For example, it is capable of annotating 52.8 million variants with allele frequencies in 6 min. AVAILABILITY: gSearch is available at http://ml.ssu.ac.kr/gSearch. It can be used as an independent search tool or can easily be integrated to existing pipelines through various programming environments such as Perl, Ruby and Python.

CompleteMOTIFs: DNA motif discovery platform for transcription factor binding experiments.

UNLABELLED: CompleteMOTIFs (cMOTIFs) is an integrated web tool developed to facilitate systematic discovery of overrepresented transcription factor binding motifs from high-throughput chromatin immunoprecipitation experiments. Comprehensive annotations and Boolean logic operations on multiple peak locations enable users to focus on genomic regions of interest for de novo motif discovery using tools such as MEME, Weeder and ChIPMunk. The pipeline incorporates a scanning tool for known motifs from TRANSFAC and JASPAR databases, and performs an enrichment test using local or precalculated background models that significantly improve the motif scanning result. Furthermore, using the cMOTIFs pipeline, we demonstrated that multiple transcription factors could cooperatively bind to the upstream of important stem cell differentiation regulators. AVAILABILITY: http://cmotifs.tchlab.org.

Taxonomizing, sizing, and overcoming the incidentalome.

PURPOSE: With the advent of whole-genome sequencing made clinically available, the number of incidental findings is likely to rise. The false-positive incidental findings are of particular clinical concern. We provide estimates on the size of these false-positive findings and classify them into four broad categories. METHODS: Whole-genome sequences (WGS) of nine individuals were scanned with several comprehensive public annotation databases and average estimates for the number of findings. These estimates were then evaluated in the perspective of various sources of false-positive annotation errors. RESULTS: At present there are four main sources of false-positive incidental findings: erroneous annotations, sequencing error, incorrect penetrance estimates, and multiple hypothesis testing. Of these, the first two are likely to be addressed in the near term. Conservatively, current methods deliver hundreds of false-positive incidental findings per individual. CONCLUSION: The burden of false-positives in whole-genome sequence interpretation threatens current capabilities to deliver clinical-grade whole-genome clinical interpretation. A new generation of population studies and retooling of the clinical decision-support approach will be required to overcome this threat.

Mapping the protein interaction network in methicillin-resistant Staphylococcus aureus.

Mortality attributable to infection with methicillin-resistant Staphylococcus aureus (MRSA) has now overtaken the death rate for AIDS in the United States, and advances in research are urgently needed to address this challenge. We report the results of the systematic identification of protein-protein interactions for the hospital-acquired strain MRSA-252. Using a high-throughput pull-down strategy combined with quantitative proteomics to distinguish specific from nonspecific interactors, we identified 13,219 interactions involving 608 MRSA proteins. Consecutive analyses revealed that this protein interaction network (PIN) exhibits scale-free organization with the characteristic presence of highly connected hub proteins. When clinical and experimental antimicrobial targets were queried in the network, they were generally found to occupy peripheral positions in the PIN with relatively few interacting partners. In contrast, the hub proteins identified in this MRSA PIN that are essential for network integrity and stability have largely been overlooked as drug targets. Thus, this empirical MRSA-252 PIN provides a rich source for identifying critical proteins essential for network stability, many of which can be considered as prospective antimicrobial drug targets.

Identification of pyruvate kinase in methicillin-resistant Staphylococcus aureus as a novel antimicrobial drug target.

Novel classes of antimicrobials are needed to address the challenge of multidrug-resistant bacteria such as methicillin-resistant Staphylococcus aureus (MRSA). Using the architecture of the MRSA interactome, we identified pyruvate kinase (PK) as a potential novel drug target based upon it being a highly connected, essential hub in the MRSA interactome. Structural modeling, including X-ray crystallography, revealed discrete features of PK in MRSA, which appeared suitable for the selective targeting of the bacterial enzyme. In silico library screening combined with functional enzymatic assays identified an acyl hydrazone-based compound (IS-130) as a potent MRSA PK inhibitor (50% inhibitory concentration [IC50] of 0.1 µM) with >1,000-fold selectivity over human PK isoforms. Medicinal chemistry around the IS-130 scaffold identified analogs that more potently and selectively inhibited MRSA PK enzymatic activity and S. aureus growth in vitro (MIC of 1 to 5 µg/ml). These novel anti-PK compounds were found to possess antistaphylococcal activity, including both MRSA and multidrug-resistant S. aureus (MDRSA) strains. These compounds also exhibited exceptional antibacterial activities against other Gram-positive genera, including enterococci and streptococci. PK lead compounds were found to be noncompetitive inhibitors and were bactericidal. In addition, mutants with significant increases in MICs were not isolated after 25 bacterial passages in culture, indicating that resistance may be slow to emerge. These findings validate the principles of network science as a powerful approach to identify novel antibacterial drug targets. They also provide a proof of principle, based upon PK in MRSA, for a research platform aimed at discovering and optimizing selective inhibitors of novel bacterial targets where human orthologs exist, as leads for anti-infective drug development.

Deep Modeling of Regulating Effects of Small Molecules on Longevity-Associated Genes.

Aging is considered an inevitable process that causes deleterious effects in the functioning and appearance of cells, tissues, and organs. Recent emergence of large-scale gene expression datasets and significant advances in machine learning techniques have enabled drug repurposing efforts in promoting longevity. In this work, we further developed our previous approach-DeepCOP, a quantitative chemogenomic model that predicts gene regulating effects, and extended its application across multiple cell lines presented in LINCS to predict aging gene regulating effects induced by small molecules. As a result, a quantitative chemogenomic Deep Model was trained using gene ontology labels, molecular fingerprints, and cell line descriptors to predict gene expression responses to chemical perturbations. Other state-of-the-art machine learning approaches were also evaluated as benchmarks. Among those, the deep neural network (DNN) classifier has top-ranked known drugs with beneficial effects on aging genes, and some of these drugs were previously shown to promote longevity, illustrating the potential utility of this methodology. These results further demonstrate the capability of "hybrid" chemogenomic models, incorporating quantitative descriptors from biomarkers to capture cell specific drug-gene interactions. Such models can therefore be used for discovering drugs with desired gene regulatory effects associated with longevity.

Drugging the 'undruggable'. Therapeutic targeting of protein-DNA interactions with the use of computer-aided drug discovery methods.

Transcription factors (TFs) act as major oncodrivers in many cancers and are frequently regarded as high-value therapeutic targets. The functionality of TFs relies on direct protein-DNA interactions, which are notoriously difficult to target with small molecules. However, this prior view of the 'undruggability' of protein-DNA interfaces has shifted substantially in recent years, in part because of significant advances in computer-aided drug discovery (CADD). In this review, we highlight recent examples of successful CADD campaigns resulting in drug candidates that directly interfere with protein-DNA interactions of several key cancer TFs, including androgen receptor (AR), ETS-related gene (ERG), MYC, thymocyte selection-associated high mobility group box protein (TOX), topoisomerase II (TOP2), and signal transducer and activator of transcription 3 (STAT3). Importantly, these findings open novel and compelling avenues for therapeutic targeting of over 1600 human TFs implicated in many conditions including and beyond cancer.

Rapid Identification of Potential Inhibitors of SARS-CoV-2 Main Protease by Deep Docking of 1.3†Billion Compounds.

The recently emerged 2019 Novel Coronavirus (SARS-CoV-2) and associated COVID-19 disease cause serious or even fatal respiratory tract infection and yet no approved therapeutics or effective treatment is currently available to effectively combat the outbreak. This urgent situation is pressing the world to respond with the development of novel vaccine or a small molecule therapeutics for SARS-CoV-2. Along these efforts, the structure of SARS-CoV-2 main protease (Mpro) has been rapidly resolved and made publicly available to facilitate global efforts to develop novel drug candidates. Recently, our group has developed a novel deep learning platform - Deep Docking (DD) which provides fast prediction of docking scores of Glide (or any other docking program) and, hence, enables structure-based virtual screening of billions of purchasable molecules in a short time. In the current study we applied DD to all 1.3 billion compounds from ZINC15 library to identify top 1,000 potential ligands for SARS-CoV-2 Mpro protein. The compounds are made publicly available for further characterization and development by scientific community.