Abstract: On screening of endolithic actinobacteria from a granite rock sample of Meghalaya for antibacterial compound, a novel antibacterial compound CCp1 was isolated from the fermentation broth of Actinomadura sp. AL2. On purification of the compound based on chromatographic techniques followed by characterization with FT-IR, UV-visible, (1)H NMR, (13)C NMR and mass spectrometry, the molecular formula of the compound was generated as C20H17N3O2, a furopyrimidine derivative. In vitro antibacterial activity of the compound was evaluated against both Gram positive and negative bacteria by agar well diffusion assay. The compound had lowest MIC (2.00 microg/ml) for Bacillus subtilis and highest MIC (> 64 microg/ml) for Staphylococcus epidermidis and Pseudomonas aeruginosa. The study revealed that the compound has potential antibacterial activity. The mode of action of the antibacterial compound was evaluated through in silico studies for its ability to bind DNA gyrase, 30S RNA molecules, OmpF porins and N-Acetylglucosamine-1-phosphate uridyltransferase (GlmU). The antibacterial compound demonstrated more favorable docking with DNA gyrase, 30S RNA molecules and OmpF porins than GlmU which support the antibacterial compound CCp1 can be as a promising broad spectrum antibiotic agent with "multitarget" characteristics.

Abstract: The emergence of large-scale genomic, chemical and pharmacological data provides new opportunities for drug discovery and repositioning. In this work, we develop a computational pipeline, called DTINet, to predict novel drug-target interactions from a constructed heterogeneous network, which integrates diverse drug-related information. DTINet focuses on learning a low-dimensional vector representation of features, which accurately explains the topological properties of individual nodes in the heterogeneous network, and then makes prediction based on these representations via a vector space projection scheme. DTINet achieves substantial performance improvement over other state-of-the-art methods for drug-target interaction prediction. Moreover, we experimentally validate the novel interactions between three drugs and the cyclooxygenase proteins predicted by DTINet, and demonstrate the new potential applications of these identified cyclooxygenase inhibitors in preventing inflammatory diseases. These results indicate that DTINet can provide a practically useful tool for integrating heterogeneous information to predict new drug-target interactions and repurpose existing drugs.Network-based data integration for drug-target prediction is a promising avenue for drug repositioning, but performance is wanting. Here, the authors introduce DTINet, whose performance is enhanced in the face of noisy, incomplete and high-dimensional biological data by learning low-dimensional vector representations.

Abstract: Protein homeostasis or proteostasis is a fundamental cellular property that encompasses the dynamic balancing of processes in the proteostasis network (PN). Such processes include protein synthesis, folding, and degradation in both non-stressed and stressful conditions. The role of the PN in neurodegenerative disease is well-documented, where it is known to respond to changes in protein folding states or toxic gain-of-function protein aggregation. Dual-specificity phosphatases have recently emerged as important participants in maintaining balance within the PN, acting through modulation of cellular signaling pathways that are involved in neurodegeneration. In this review, we will summarize recent findings describing the roles of dual-specificity phosphatases in neurodegeneration and offer perspectives on future therapeutic directions.

Abstract: BACKGROUND: Dispersed biomedical databases limit user exploration to generate structured knowledge. Linked Data unifies data structures and makes the dispersed data easy to search across resources, but it lacks supporting human cognition to achieve insights. In addition, potential errors in the data are difficult to detect in their free formats. Devising a visualization that synthesizes multiple sources in such a way that links between data sources are transparent, and uncertainties, such as data conflicts, are salient is challenging. RESULTS: To investigate the requirements and challenges of uncertainty-aware visualizations of linked data, we developed MediSyn, a system that synthesizes medical datasets to support drug treatment selection. It uses a matrix-based layout to visually link drugs, targets (e.g., mutations), and tumor types. Data uncertainties are salient in MediSyn; for example, (i) missing data are exposed in the matrix view of drug-target relations; (ii) inconsistencies between datasets are shown via overlaid layers; and (iii) data credibility is conveyed through links to data provenance. CONCLUSIONS: Through the synthesis of two manually curated datasets, cancer treatment biomarkers and drug-target bioactivities, a use case shows how MediSyn effectively supports the discovery of drug-repurposing opportunities. A study with six domain experts indicated that MediSyn benefited the drug selection and data inconsistency discovery. Though linked publication sources supported user exploration for further information, the causes of inconsistencies were not easy to find. Additionally, MediSyn could embrace more patient data to increase its informativeness. We derive design implications from the findings.

Abstract: Analysis of drug-target interactions (DTIs) is of great importance in developing new drug candidates for known protein targets or discovering new targets for old drugs. However, the experimental approaches for identifying DTIs are expensive, laborious and challenging. In this study, we report a novel computational method for predicting DTIs using the highly discriminative information of drug-target interactions and our newly developed discriminative vector machine (DVM) classifier. More specifically, each target protein sequence is transformed as the position-specific scoring matrix (PSSM), in which the evolutionary information is retained; then the local binary pattern (LBP) operator is used to calculate the LBP histogram descriptor. For a drug molecule, a novel fingerprint representation is utilized to describe its chemical structure information representing existence of certain functional groups or fragments. When applying the proposed method to the four datasets (Enzyme, GPCR, Ion Channel and Nuclear Receptor) for predicting DTIs, we obtained good average accuracies of 93.16%, 89.37%, 91.73% and 92.22%, respectively. Furthermore, we compared the performance of the proposed model with that of the state-of-the-art SVM model and other previous methods. The achieved results demonstrate that our method is effective and robust and can be taken as a useful tool for predicting DTIs.

Abstract: Neurons are often classified by their morphological and molecular properties. The online knowledge base Hippocampome.org primarily defines neuron types from the rodent hippocampal formation based on their main neurotransmitter (glutamate or GABA) and the spatial distributions of their axons and dendrites. For each neuron type, this open-access resource reports any and all published information regarding the presence or absence of known molecular markers, including calcium-binding proteins, neuropeptides, receptors, channels, transcription factors, and other molecules of biomedical relevance. The resulting chemical profile is relatively sparse: even for the best studied neuron types, the expression or lack thereof of fewer than 70 molecules has been firmly established to date. The mouse genome-wide in situ hybridization mapping of the Allen Brain Atlas provides a wealth of data that, when appropriately analyzed, can substantially augment the molecular marker knowledge in Hippocampome.org. Here we focus on the principal cell layers of dentate gyrus (DG), CA3, CA2, and CA1, which together contain approximately 90% of hippocampal neurons. These four anatomical parcels are densely packed with somata of mostly excitatory projection neurons. Thus, gene expression data for those layers can be justifiably linked to the respective principal neuron types: granule cells in DG and pyramidal cells in CA3, CA2, and CA1. In order to enable consistent interpretation across genes and regions, we screened the whole-genome dataset against known molecular markers of those neuron types. The resulting threshold values allow over 6000 very-high confidence (>99.5%) expressed/not-expressed assignments, expanding the biochemical information content of Hippocampome.org more than five-fold. Many of these newly identified molecular markers are potential pharmacological targets for major neurological and psychiatric conditions. Furthermore, our approach yields reasonable expression/non-expression estimates for every single gene in each of these four neuron types with >90% average confidence, providing a considerably complete genetic characterization of hippocampal principal neurons.

Abstract: This is a comparative pharmacokinetics study of linezolid (Lzd), and two novel oxazolidinone antibacterial agents-PH027 and PH051-in rabbits to determine if the discrepancy between the in vitro and in vivo activities of the novel compounds is due to pharmacokinetic factors. The pharmacokinetics after IV and oral administration, plasma protein binding and tissue distribution for the three compounds were compared. The elimination half-lives were 52.4 +/- 6.3, 68.7 +/- 12.1 and 175 +/- 46.1 min for Lzd, PH027 and PH051, respectively. The oral bioavailability for Lzd, PH027 and PH051 administered as suspension were 38.7%, 22.1% and 4.73%, which increased significantly when administered as microemulsion to 51.7%, 72.9% and 13.9%. The plasma protein binding were 32-34%, 37-38% and 90-91% for Lzd, PH027 and PH051. The tissue distribution for PH027 and PH051 in all investigated tissues were higher than that for Lzd. It can be concluded that the lower bioavailability of PH027 and PH051 compared to Lzd when administered as suspension is the main cause of their lower in vivo activity, despite their comparable in vitro activity. Differences in the other pharmacokinetic characteristics cannot explain the lower in vivo activity. The in vivo activity of the novel compounds should be re-evaluated using formulations with good oral bioavailability.

Abstract: Potential drug target identification and mechanism of action is an important step in drug discovery process, which can be achieved by biochemical methods, genetic interactions or computational conjectures. Sometimes more than one approach is implemented to mine out the potential drug target and characterize the on-target or off-target effects. A novel anticancer agent RH1 is designed as pro-drug to be activated by NQO1, an enzyme overexpressed in many types of tumors. However, increasing data show that RH1 can affect cells in NQO1-independent fashion. Here, we implemented the bioinformatics approach of modeling and molecular docking for search of RH1 targets among protein kinase species. We have examined 129 protein kinases in total where 96 protein kinases are in complexes with their inhibitor, 11 kinases were in the unbound state with any ligand and for 22 protein kinases 3D structure were modeled. Comparison of calculated free energy of binding of RH1 with indigenous kinase inhibitors binding efficiency as well as alignment of their pharmacophoric maps let us predict and ranked protein kinases such as KIT, CDK2, CDK6, MAPK1, NEK2 and others as the most prominent off-targets of RH1. Our finding opens new avenues in search of protein targets that might be responsible for curing cancer by new promising drug RH1 in NQO1-independent way.

Abstract: Control theory is a well-established approach in network science, with applications in bio-medicine and cancer research. We build on recent results for structural controllability of directed networks, which identifies a set of driver nodes able to control an a-priori defined part of the network. We develop a novel and efficient approach for the (targeted) structural controllability of cancer networks and demonstrate it for the analysis of breast, pancreatic, and ovarian cancer. We build in each case a protein-protein interaction network and focus on the survivability-essential proteins specific to each cancer type. We show that these essential proteins are efficiently controllable from a relatively small computable set of driver nodes. Moreover, we adjust the method to find the driver nodes among FDA-approved drug-target nodes. We find that, while many of the drugs acting on the driver nodes are part of known cancer therapies, some of them are not used for the cancer types analyzed here; some drug-target driver nodes identified by our algorithms are not known to be used in any cancer therapy. Overall we show that a better understanding of the control dynamics of cancer through computational modelling can pave the way for new efficient therapeutic approaches and personalized medicine.

Abstract: Drug-drug interactions (DDIs) are known to be responsible for nearly a third of all adverse drug reactions. Hence several current efforts focus on extracting signal from EMRs to prioritize DDIs that need further exploration. To this end, being able to extract explicit mentions of DDIs in free text narratives is an important task. In this paper, we explore recurrent neural network (RNN) architectures to detect and classify DDIs from unstructured text using the DDIExtraction dataset from the SemEval 2013 (task 9) shared task. Our methods are in line with those used in other recent deep learning efforts for relation extraction including DDI extraction. However, to our knowledge, we are the first to investigate the potential of character-level RNNs (Char-RNNs) for DDI extraction (and relation extraction in general). Furthermore, we explore a simple but effective model bootstrapping method to (a). build model averaging ensembles, (b). derive confidence intervals around mean micro-F scores (MMF), and (c). assess the average behavior of our methods. Without any rule based filtering of negative examples, a popular heuristic used by most earlier efforts, we achieve an MMF of 69.13. By adding simple replicable heuristics to filter negative instances we are able to achieve an MMF of 70.38. Furthermore, our best ensembles produce micro F-scores of 70.81 (without filtering) and 72.13 (with filtering), which are superior to metrics reported in published results. Although Char-RNNs turnout to be inferior to regular word based RNN models in overall comparisons, we find that ensembling models from both architectures results in nontrivial gains over simply using either alone, indicating that they complement each other.