Drug-Drug Interaction Extraction via Recurrent Hybrid Convolutional Neural Networks with an Improved Focal Loss.

Drug-drug interactions (DDIs) may bring huge health risks and dangerous effects to a patient's body when taking two or more drugs at the same time or within a certain period of time. Therefore, the automatic extraction of unknown DDIs has great potential for the development of pharmaceutical agents and the safety of drug use. In this article, we propose a novel recurrent hybrid convolutional neural network (RHCNN) for DDI extraction from biomedical literature. In the embedding layer, the texts mentioning two entities are represented as a sequence of semantic embeddings and position embeddings. In particular, the complete semantic embedding is obtained by the information fusion between a word embedding and its contextual information which is learnt by recurrent structure. After that, the hybrid convolutional neural network is employed to learn the sentence-level features which consist of the local context features from consecutive words and the dependency features between separated words for DDI extraction. Lastly but most significantly, in order to make up for the defects of the traditional cross-entropy loss function when dealing with class imbalanced data, we apply an improved focal loss function to mitigate against this problem when using the DDIExtraction 2013 dataset. In our experiments, we achieve DDI automatic extraction with a micro F-score of 75.48% on the DDIExtraction 2013 dataset, outperforming the state-of-the-art approach by 2.49%.

A water-soluble near-infrared fluorescent probe for specific Pd2+ detection.

Palladium (Pd) is widely used in chemistry, biology, environmental science etc., and Pd2+ is the most plenitudinous oxidation state of the Pd that can exist under physiological conditions or in living cells, which could have adverse effects on both our health and environment. Thus, it is of great significance to monitor the changes of Pd2+. Hence, a novel near-infrared fluorescent probe M-PD has been developed for selective detection of Pd2+ based on naphthofluorescein in this work. The result demonstrated that M-PD exhibited favorable properties for sensing Pd2+ such as excellent water solubility, high selectivity and sensitivity. And the limit of detection was estimated as 10.8 nM, much lower than the threshold in drugs (5-10 ppm) specified by European Directorate for the Quality Control of Medicines. More importantly, detection and recovery experiments of Pd2+ in aspirin aqoeous solution and soil are satisfactory. In addition, M-PD has also been successfully used for near-infrared fluorescence imaging of Pd2+ in living cells, indicating that the probe has better feasibility and application potential in the determination of Pd2+.

Systems pharmacology dissection of the anti-inflammatory mechanism for the medicinal herb Folium eriobotryae.

Inflammation is a hallmark of many diseases like diabetes, cancers, atherosclerosis and arthritis. Thus, lots of concerns have been raised toward developing novel anti-inflammatory agents. Many alternative herbal medicines possess excellent anti-inflammatory properties, yet their precise mechanisms of action are yet to be elucidated. Here, a novel systems pharmacology approach based on a large number of chemical, biological and pharmacological data was developed and exemplified by a probe herb Folium Eriobotryae, a widely used clinical anti-inflammatory botanic drug. The results show that 11 ingredients of this herb with favorable pharmacokinetic properties are predicted as active compounds for anti-inflammatory treatment. In addition, via systematic network analyses, their targets are identified to be 43 inflammation-associated proteins including especially COX2, ALOX5, PPARG, TNF and RELA that are mainly involved in the mitogen-activated protein kinase (MAPK) signaling pathway, the rheumatoid arthritis pathway and NF-κB signaling pathway. All these demonstrate that the integrated systems pharmacology method provides not only an effective tool to illustrate the anti-inflammatory mechanisms of herbs, but also a new systems-based approach for drug discovery from, but not limited to, herbs, especially when combined with further experimental validations.

The tendency to recreate ancestral CG dinucleotides in the human genome.

BACKGROUND: The CG dinucleotides are known to be deficient in the human genome, due to a high mutation rate from 5-methylated CG to TG and its complementary pair CA. Meanwhile, many cellular functions rely on these CG dinucleotides, such as gene expression controlled by cytosine methylation status. Thus, CG dinucleotides that provide essential functional substrates should be retained in genomes. How these two conflicting processes regarding the fate of CG dinucleotides - i.e., high mutation rate destroying CG dinucleotides, vs. functional processes that require their preservation remains an unsolved question. RESULTS: By analyzing the mutation and frequency spectrum of newly derived alleles in the human genome, a tendency towards generating more CGs was observed, which was mainly contributed by an excess number of mutations from CA/TG to CG. Simultaneously, we found a fixation preference for CGs derived from TG/CA rather than CGs generated by other dinucleotides. These tendencies were observed both in intergenic and genic regions. An analysis of Integrated Extended Haplotype Homozygosity provided no evidence of selection for newly derived CGs. CONCLUSIONS: Ancestral CG dinucleotides that were subsequently lost by mutation tend to be recreated in the human genome, as indicated by a biased mutation and fixation pattern favoring new CGs that derived from TG/CA.

DYNAMIK: a software environment for cell DYNAmics, Motility, and Information tracKing, with an application to Ras pathways.

The emergence of new microscopy techniques in combination with the increasing resource of bioimaging data has given fresh impetus to utilizing image processing methods for studying biological processes. Cell tracking studies in particular, which are important for a wide range of biological processes such as embryonic development or the immune system, have recently become the focus of attention. These studies typically produce large volumes of data that are hard to investigate manually and therefore call for an automated approach. Due to the large variety of biological cells and the inhomogeneity of applications, however, there exists no widely accepted method or system for cell tracking until today. In this article, we present our publicly available DYNAMIK software environment that allows users to compute a suit of cell features and plot the trajectory of multiple cells over a sequence of frames. Using chemotaxis and Ras pathways as an example, we show how users can employ our software to compute statistics about cell motility and other cell information, and how to evaluate their test series based on the data computed. We see that DYNAMIK's segmentation and tracking compares favorably with the output produced by other software packages.

Specificity rule discovery in HIV-1 protease cleavage site analysis.

Several machine learning algorithms have recently been applied to modeling the specificity of HIV-1 protease. The problem is challenging because of the three issues as follows: (1) datasets with high dimensionality and small number of samples could misguide classification modeling and its interpretation; (2) symbolic interpretation is desirable because it provides us insight to the specificity in the form of human-understandable rules, and thus helps us to design effective HIV inhibitors; (3) the interpretation should take into account complexity or dependency between positions in sequences. Therefore, it is necessary to investigate multivariate and feature-selective methods to model the specificity and to extract rules from the model. We have tested extensively various machine learning methods, and we have found that the combination of neural networks and decompositional approach can generate a set of effective rules. By validation to experimental results for the HIV-1 protease, the specificity rules outperform the ones generated by frequency-based, univariate or black-box methods.

A luminescent bimetallic iridium(iii) complex for ratiometric tracking intracellular viscosity.

A luminescent bimetallic iridium probe C10 was developed through a long soft carbon chain linkage to achieve ratiometric detection of viscosity. C10 features high sensitivity and selectivity for viscosity. More importantly, C10 is living cell permeable and can be employed to distinguish cancer cells from normal cells and track viscosity changes during MCF-7 cell apoptosis.

Correction: A luminescent bimetallic iridium(iii) complex for ratiometric tracking intracellular viscosity.

Correction for 'A luminescent bimetallic iridium(iii) complex for ratiometric tracking intracellular viscosity' by Fengyu Liu et al., Chem. Commun., 2018, 54, 1371-1374.

A Methodology for Cancer Therapeutics by Systems Pharmacology-Based Analysis: A Case Study on Breast Cancer-Related Traditional Chinese Medicines.

Breast cancer is the most common carcinoma in women. Comprehensive therapy on breast cancer including surgical operation, chemotherapy, radiotherapy, endocrinotherapy, etc. could help, but still has serious side effect and resistance against anticancer drugs. Complementary and alternative medicine (CAM) may avoid these problems, in which traditional Chinese medicine (TCM) has been highlighted. In this section, to analyze the mechanism through which TCM act on breast cancer, we have built a virtual model consisting of the construction of database, oral bioavailability prediction, drug-likeness evaluation, target prediction, network construction. The 20 commonly employed herbs for the treatment of breast cancer were used as a database to carry out research. As a result, 150 ingredient compounds were screened out as active molecules for the herbs, with 33 target proteins predicted. Our analysis indicates that these herbs 1) takes a 'Jun-Chen-Zuo-Shi" as rule of prescription, 2) which function mainly through perturbing three pathways involving the epidermal growth factor receptor, estrogen receptor, and inflammatory pathways, to 3) display the breast cancer-related anti-estrogen, anti-inflammatory, regulation of cell metabolism and proliferation activities. To sum it up, by providing a novel in silico strategy for investigation of the botanical drugs, this work may be of some help for understanding the action mechanisms of herbal medicines and for discovery of new drugs from plants.

State of the Science: Cancer Complementary and Alternative Medicine Therapeutics Research-NCI Strategic Workshop Highlights of Discussion Report.

In May 2016, the Office of Cancer Complementary and Alternative Medicine, Division of Cancer Diagnosis and Treatment, of the National Cancer Institute convened a special workshop focused on the State of the Science: Cancer Complementary and Alternative Medicine Therapeutics Research. The current state of the science, gaps, and future opportunities were reviewed and discussed by a distinguished panel of experts in this field of research, and the highlights of this meeting are reported herein.

Statistical distributions of optimal global alignment scores of random protein sequences.

BACKGROUND: The inference of homology from statistically significant sequence similarity is a central issue in sequence alignments. So far the statistical distribution function underlying the optimal global alignments has not been completely determined. RESULTS: In this study, random and real but unrelated sequences prepared in six different ways were selected as reference datasets to obtain their respective statistical distributions of global alignment scores. All alignments were carried out with the Needleman-Wunsch algorithm and optimal scores were fitted to the Gumbel, normal and gamma distributions respectively. The three-parameter gamma distribution performs the best as the theoretical distribution function of global alignment scores, as it agrees perfectly well with the distribution of alignment scores. The normal distribution also agrees well with the score distribution frequencies when the shape parameter of the gamma distribution is sufficiently large, for this is the scenario when the normal distribution can be viewed as an approximation of the gamma distribution. CONCLUSION: We have shown that the optimal global alignment scores of random protein sequences fit the three-parameter gamma distribution function. This would be useful for the inference of homology between sequences whose relationship is unknown, through the evaluation of gamma distribution significance between sequences.

A systems pharmacology approach to investigate the mechanisms of action of Semen Strychni and Tripterygium wilfordii Hook F for treatment of rheumatoid arthritis.

ETHNOPHARMACOLOGY RELEVANCE: The angiogenesis control at the initiation of rheumatoid arthritis (RA) that mainly blocks the inflammatory cascades expects to attenuate the action of angiogenic mediators, synovial angiogenesis, and to partially reverse the erosive bone damage. Two typical Chinese herbs, Semen Strychni and Tripterygium wilfordii Hook F (TwHF) have been used as a remedy to treat RA since ancient time. However, their functioning mechanisms are still unknown. Thus it is necessary to exploit their underlying mechanism for the treatment of RA. METHODS: This study was undertaken to analyze their underlying mechanism based on a systems biology platform. Firstly, active components of the two herbs were screened out from TcmSP database based on their OB and DL values. Then their potential targets were predicted by using Random Forest, Support Vector Machine, and validated via docking process. Finally, a network of compound-target was constructed. RESULTS: In this work, 27 and 33 active compounds were screened out from Semen Strychni and TwHF, targeting 28 and 32 potential proteins, respectively. The results show that the two herbs modulate the angiogenesis mediators through both direct and indirect pathways, and 21 common targets shared by Semen Strychni and TwHF bear major responsibility for treating RA. CONCLUSIONS: The underlying mechanism of Semen Strychni and TwHF in treatment of RA is through multiple targets interaction by their blocking of the angiogenesis mediator cascades. This may provide us a better understanding of the function of the two herbs for the treatment of RA, as well as a clue to unveil their possible treatment effects of other systemic diseases, and in this way, hopefully the screening models may facilitate the discovery of novel combined drugs.

Systems pharmacology to decipher the combinational anti-migraine effects of Tianshu formula.

Migraine is the most common neurovascular disorder that imparts a considerable burden to health care system around the world. However, currently there are still no effective and widely applicable pharmacotherapies for migraine patients. Herbal formulae, characterized as multiple herbs, constituents and targets, have been acknowledged with clinical effects in treating migraine, which attract more and more researchers' attention although their exact molecular mechanisms are still unclear. In this work, a novel systems pharmacology-based method which integrates pharmacokinetic filtering, target fishing and network analysis was developed and exemplified by a probe, i.e. Tianshu formula, a widely clinically used anti-migraine herbal formula in China which comprises of Rhizoma chuanxiong and Gastrodia elata. The results exhibit that 20 active ingredients of Tianshu formula possess favorable pharmacokinetic profiles, which have interactions with 48 migraine-related targets to provide potential synergistic therapeutic effects. Additionally, from systematic analysis, we speculate that R. chuanxiong as the monarch herb mediates the major targets like PTGS2, ESR1, NOS2, HTR1B and NOS3 to regulate the vascular and nervous systems, as well as the inflammation and pain-related pathways to benefit migraine patients. Meanwhile, as an adjuvant herb, G. elata may not only assist the monarch herb to improve the outcome of migraine patients, but also regulate multiple targets like ABAT, HTR1D, ALOX15 and KCND3 to modify migraine accompanying symptoms like vomiting, vertigo and gastrointestinal disorders.

A diagnostic methodology for Alzheimer's disease.

BACKGROUND: Like all other neurodegenerative diseases, Alzheimer's disease (AD) remains a very challenging and difficult problem for diagnosis and therapy. For many years, only historical, behavioral and psychiatric measures have been available to AD cases. Recently, a definitive diagnostic framework, using biomarkers and imaging, has been proposed. In this paper, we propose a promising diagnostic methodology for the framework. METHODS: In a previous paper, we developed an efficient SVM (Support Vector Machine) based method, which we have now applied to discover important biomarkers and target networks which provide strategies for AD therapy. RESULTS: The methodology selects a number of blood-based biomarkers (fewer than 10% of initial numbers on three AD datasets from NCBI), and the results are statistically verified by cross-validation. The resulting SVM is a classifier of AD vs. normal subjects. We construct target networks of AD based on MI (mutual information). In addition, a hierarchical clustering is applied on the initial data and clustered genes are visualized in a heatmap. The proposed method also performs gender analysis by classifying subjects based on gender. CONCLUSIONS: Unlike other traditional statistical analyses, our method uses a machine learning-based algorithm. Our method selects a small set of important biomarkers for AD, differentiates noisy (irrelevant) from relevant biomarkers and also provides the target networks of the selected biomarkers, which will be useful for diagnosis and therapeutic design. Finally, based on the gender analysis, we observe that gender could play a role in AD diagnosis.

Cancer classification: Mutual information, target network and strategies of therapy.

BACKGROUND: Cancer therapy is a challenging research area because side effects often occur in chemo and radiation therapy. We intend to study a multi-targets and multi-components design that will provide synergistic results to improve efficiency of cancer therapy. METHODS: We have developed a general methodology, AMFES (Adaptive Multiple FEature Selection), for ranking and selecting important cancer biomarkers based on SVM (Support Vector Machine) classification. In particular, we exemplify this method by three datasets: a prostate cancer (three stages), a breast cancer (four subtypes), and another prostate cancer (normal vs. cancerous). Moreover, we have computed the target networks of these biomarkers as the signatures of the cancers with additional information (mutual information between biomarkers of the network). Then, we proposed a robust framework for synergistic therapy design approach which includes varies existing mechanisms. RESULTS: These methodologies were applied to three GEO datasets: GSE18655 (three prostate stages), GSE19536 (4 subtypes breast cancers) and GSE21036 (prostate cancer cells and normal cells) shown in. We selected 96 biomarkers for first prostate cancer dataset (three prostate stages), 72 for breast cancer (luminal A vs. luminal B), 68 for breast cancer (basal-like vs. normal-like), and 22 for another prostate cancer (cancerous vs. normal. In addition, we obtained statistically significant results of mutual information, which demonstrate that the dependencies among these biomarkers can be positive or negative. CONCLUSIONS: We proposed an efficient feature ranking and selection scheme, AMFES, to select an important subset from a large number of features for any cancer dataset. Thus, we obtained the signatures of these cancers by building their target networks. Finally, we proposed a robust framework of synergistic therapy for cancer patients. Our framework is not only supported by real GEO datasets but also aim to a multi-targets/multi-components drug design tool, which improves the traditional single target/single component analysis methods. This framework builds a computational foundation which can provide a clear classification of cancers and lead to an efficient cancer therapy.

Hierarchical modularity in ERα transcriptional network is associated with distinct functions and implicates clinical outcomes.

Recent genome-wide profiling reveals highly complex regulation networks among ERα and its targets. We integrated estrogen (E2)-stimulated time-series ERα ChIP-seq and gene expression data to identify the ERα-centered transcription factor (TF) hubs and their target genes, and inferred the time-variant hierarchical network structures using a Bayesian multivariate modeling approach. With its recurrent motif patterns, we determined three embedded regulatory modules from the ERα core transcriptional network. The GO analyses revealed the distinct biological function associated with each of three embedded modules. The survival analysis showed the genes in each module were able to render a significant survival correlation in breast cancer patient cohorts. In summary, our Bayesian statistical modeling and modularity analysis not only reveals the dynamic properties of the ERα-centered regulatory network and associated distinct biological functions, but also provides a reliable and effective genomic analytical approach for the analysis of dynamic regulatory network for any given TF.

Integrative identification of core genetic regulatory modules via a structural model-based clustering method.

Regulatory modules play fundamental roles in processing and dispatching signals in cell life cycle. Although current clustering methods may reduce data complexity to lower dimension, they tend to neglect biological meanings within high-throughput data. We propose a module-detection algorithm through defining network activity measures and associating them through a weighted clustering approach. We verify our method on diverse models and it provides a unique perspective for analysing model dynamics and expression data, especially with consideration of inherent biological meanings. As it can detect core regulatory modules effectively, it facilitates pathway/network modelling in systems biology.

Computational inference and analysis of genetic regulatory networks via a supervised combinatorial-optimization pattern.

BACKGROUND: Post-genome era brings about diverse categories of omics data. Inference and analysis of genetic regulatory networks act prominently in extracting inherent mechanisms, discovering and interpreting the related biological nature and living principles beneath mazy phenomena, and eventually promoting the well-beings of humankind. RESULTS: A supervised combinatorial-optimization pattern based on information and signal-processing theories is introduced into the inference and analysis of genetic regulatory networks. An associativity measure is proposed to define the regulatory strength/connectivity, and a phase-shift metric determines regulatory directions among components of the reconstructed networks. Thus, it solves the undirected regulatory problems arising from most of current linear/nonlinear relevance methods. In case of computational and topological redundancy, we constrain the classified group size of pair candidates within a multiobjective combinatorial optimization (MOCO) pattern. CONCLUSIONS: We testify the proposed approach on two real-world microarray datasets of different statistical characteristics. Thus, we reveal the inherent design mechanisms for genetic networks by quantitative means, facilitating further theoretic analysis and experimental design with diverse research purposes. Qualitative comparisons with other methods and certain related focuses needing further work are illustrated within the discussion section.

Text mining biomedical literature for constructing gene regulatory networks.

In this paper, we present the framework of a Gene Regulatory Networks System: GRNS. The goals of GRNS include automatically mining biomedical literature to extract gene regulatory information (strain number, genotype, gene regulatory relation, and phenotype), automatically constructing gene regulatory networks based on extracted information and integrating biomedical knowledge into the regulatory networks.

Translational systems genomics: ontology and imaging.

In developing an integrated framework for translational bioinformatics, we consider bioimaging in the NIH Roadmap that exploits high-resolution genomic imaging for clinical applications to the diagnosis and treatment of genetic disorders/diseases. On one hand, we develop new image processing techniques, while on the other, we use the fusion of several well known ontological standards - Gene Ontology (GO), Clinical Bioinformatics Ontology (CBO), Foundational Model of Anatomy (FMA) and Microarry Gene Expression Data Ontology (MGED) in this framework. We have discovered that the heterogeneity of the imaging data can be resolved at the different ontological levels of this framework. Moreover, structural genomic information can be readily integrated into the usual textual clinical information bases.