Drug-drug relationship based on target information: application to drug target identification.
ABSTRACT: BACKGROUND: Drugs that bind to common targets likely exert similar activities. In this target-centric view, the inclusion of richer target information may better represent the relationships between drugs and their activities. Under this assumption, we expanded the "common binding rule" assumption of QSAR to create a new drug-drug relationship score (DRS). METHOD: Our method uses various chemical features to encode drug target information into the drug-drug relationship information. Specifically, drug pairs were transformed into numerical vectors containing the basal drug properties and their differences. After that, machine learning techniques such as data cleaning, dimension reduction, and ensemble classifier were used to prioritize drug pairs bound to a common target. In other words, the estimation of the drug-drug relationship is restated as a large-scale classification problem, which provides the framework for using state-of-the-art machine learning techniques with thousands of chemical features for newly defining drug-drug relationships. CONCLUSIONS: Various aspects of the presented score were examined to determine its reliability and usefulness: the abundance of common domains for the predicted drug pairs, c.a. 80% coverage for known targets, successful identifications of unknown targets, and a meaningful correlation with another cutting-edge method for analyzing drug similarities. The most significant strength of our method is that the DRS can be used to describe phenotypic similarities, such as pharmacological effects.
Project description:Tuberculosis (TB) remains a significant global health problem for which rapid diagnosis is critical to both treatment and control. This report describes a multiplex PCR method, the Mycobacterial IDentification and Drug Resistance Screen (MID-DRS) assay, which allows identification of members of the Mycobacterium tuberculosis complex (MTBC) and the simultaneous amplification of targets for sequencing-based drug resistance screening of rifampin-resistant (rifampin(r)), isoniazid(r), and pyrazinamide(r) TB. Additionally, the same multiplex reaction amplifies a specific 16S rRNA gene target for rapid identification of M. avium complex (MAC) and a region of the heat shock protein 65 gene (hsp65) for further DNA sequencing-based confirmation or identification of other mycobacterial species. Comparison of preliminary results generated with MID-DRS versus culture-based methods for a total of 188 bacterial isolates demonstrated MID-DRS sensitivity and specificity as 100% and 96.8% for MTBC identification; 100% and 98.3% for MAC identification; 97.4% and 98.7% for rifampin(r) TB identification; 60.6% and 100% for isoniazid(r) TB identification; and 75.0% and 98.1% for pyrazinamide(r) TB identification. The performance of the MID-DRS was also tested on acid-fast-bacterium (AFB)-positive clinical specimens, resulting in sensitivity and specificity of 100% and 78.6% for detection of MTBC and 100% and 97.8% for detection of MAC. In conclusion, use of the MID-DRS reduces the time necessary for initial identification and drug resistance screening of TB specimens to as little as 2 days. Since all targets needed for completing the assay are included in a single PCR amplification step, assay costs, preparation time, and risks due to user errors are also reduced.
Project description:A current key feature in drug-target network is that drugs often bind to multiple targets, known as polypharmacology or drug promiscuity. Recent literature has indicated that relatively small fragments in both drugs and targets are crucial in forming polypharmacology. We hypothesize that principles behind polypharmacology are embedded in paired fragments in molecular graphs and amino acid sequences of drug-target interactions. We developed a fast, scalable algorithm for mining significantly co-occurring subgraph-subsequence pairs from drug-target interactions. A noteworthy feature of our approach is to capture significant paired patterns of subgraph-subsequence, while patterns of either drugs or targets only have been considered in the literature so far. Significant substructure pairs allow the grouping of drug-target interactions into clusters, covering approximately 75% of interactions containing approved drugs. These clusters were highly exclusive to each other, being statistically significant and logically implying that each cluster corresponds to a distinguished type of polypharmacology. These exclusive clusters cannot be easily obtained by using either drug or target information only but are naturally found by highlighting significant substructure pairs in drug-target interactions. These results confirm the effectiveness of our method for interpreting polypharmacology in drug-target network.
Project description:A common drug repositioning strategy is the re-application of an existing drug to address alternative targets. A crucial aspect to enable such repurposing is that the drug's binding site on the original target is similar to that on the alternative target. Based on the assumption that proteins with similar binding sites may bind to similar drugs, the 3D substructure similarity data can be used to identify similar sites in other proteins that are not known targets. The Drug ReposER (DRug REPOSitioning Exploration Resource) web server is designed to identify potential targets for drug repurposing based on sub-structural similarity to the binding interfaces of known drug binding sites. The application has pre-computed amino acid arrangements from protein structures in the Protein Data Bank that are similar to the 3D arrangements of known drug binding sites thus allowing users to explore them as alternative targets. Users can annotate new structures for sites that are similarly arranged to the residues found in known drug binding interfaces. The search results are presented as mappings of matched sidechain superpositions. The results of the searches can be visualized using an integrated NGL viewer. The Drug ReposER server has no access restrictions and is available at http://mfrlab.org/drugreposer/.
Project description:Adverse drug effects (ADEs) are one of the leading causes of death in developed countries and are the main reason for drug recalls from the market, whereas the ADEs that are associated with action on the cardiovascular system are the most dangerous and widespread. The treatment of human diseases often requires the intake of several drugs, which can lead to undesirable drug-drug interactions (DDIs), thus causing an increase in the frequency and severity of ADEs. An evaluation of DDI-induced ADEs is a nontrivial task and requires numerous experimental and clinical studies. Therefore, we developed a computational approach to assess the cardiovascular ADEs of DDIs. This approach is based on the combined analysis of spontaneous reports (SRs) and predicted drug-target interactions to estimate the five cardiovascular ADEs that are induced by DDIs, namely, myocardial infarction, ischemic stroke, ventricular tachycardia, cardiac failure, and arterial hypertension. We applied a method based on least absolute shrinkage and selection operator (LASSO) logistic regression to SRs for the identification of interacting pairs of drugs causing corresponding ADEs, as well as noninteracting pairs of drugs. As a result, five datasets containing, on average, 3100 potentially ADE-causing and non-ADE-causing drug pairs were created. The obtained data, along with information on the interaction of drugs with 1553 human targets predicted by PASS Targets software, were used to create five classification models using the Random Forest method. The average area under the ROC curve of the obtained models, sensitivity, specificity and balanced accuracy were 0.837, 0.764, 0.754 and 0.759, respectively. The predicted drug targets were also used to hypothesize the potential mechanisms of DDI-induced ventricular tachycardia for the top-scoring drug pairs. The created five classification models can be used for the identification of drug combinations that are potentially the most or least dangerous for the cardiovascular system.
Project description:As pharmacodynamic drug-drug interactions (PD DDIs) could lead to severe adverse effects in patients, it is important to identify potential PD DDIs in drug development. The signaling starting from drug targets is propagated through protein-protein interaction (PPI) networks. PD DDIs could occur by close interference on the same targets or within the same pathways as well as distant interference through cross-talking pathways. However, most of the previous approaches have considered only close interference by measuring distances between drug targets or comparing target neighbors. We have applied a random walk with restart algorithm to simulate signaling propagation from drug targets in order to capture the possibility of their distant interference. Cross validation with DrugBank and Kyoto Encyclopedia of Genes and Genomes DRUG shows that the proposed method outperforms the previous methods significantly. We also provide a web service with which PD DDIs for drug pairs can be analyzed at http://biosoft.kaist.ac.kr/targetrw.
Project description:In pharmaceutical sciences, a crucial step of the drug discovery process is the identification of drug-target interactions. However, only a small portion of the drug-target interactions have been experimentally validated, as the experimental validation is laborious and costly. To improve the drug discovery efficiency, there is a great need for the development of accurate computational approaches that can predict potential drug-target interactions to direct the experimental verification. In this paper, we propose a novel drug-target interaction prediction algorithm, namely neighborhood regularized logistic matrix factorization (NRLMF). Specifically, the proposed NRLMF method focuses on modeling the probability that a drug would interact with a target by logistic matrix factorization, where the properties of drugs and targets are represented by drug-specific and target-specific latent vectors, respectively. Moreover, NRLMF assigns higher importance levels to positive observations (i.e., the observed interacting drug-target pairs) than negative observations (i.e., the unknown pairs). Because the positive observations are already experimentally verified, they are usually more trustworthy. Furthermore, the local structure of the drug-target interaction data has also been exploited via neighborhood regularization to achieve better prediction accuracy. We conducted extensive experiments over four benchmark datasets, and NRLMF demonstrated its effectiveness compared with five state-of-the-art approaches.
Project description:BACKGROUND:For treating a complex disease such as cancer, some effective means are needed to control biological networks that underlies the disease. The one-target one-drug paradigm has been the dominating drug discovery approach in the past decades. Compared to single target-based drugs, combination drug targets may overcome many limitations of single drug target and achieve a more effective and safer control of the disease. Most of existing combination drug targets are developed based on clinical experience or text-and-trial strategy, which cannot provide theoretical guidelines for designing and screening effective drug combinations. Therefore, systematic identification of multiple drug targets and optimal intervention strategy needs to be developed. RESULTS:We developed a strategy to screen the synergistic combinations of two drug targets in disease networks based on the classification of single drug targets. The method tried to identify the sensitivity of single intervention and then the combination of multiple interventions that can restore the disease network to a desired normal state. In our strategy of screening drug target combinations, we first classified all drug targets into sensitive and insensitive single drug targets. Then, we identified the synergistic and antagonistic of drug target combinations, including the combinations of sensitive drug targets, the combinations of insensitive drug target and the combination of sensitive and insensitive targets. Finally, we applied our strategy to Arachidonic Acid (AA) metabolic network and found 18 pairs of synergistic drug target combinations, five of which have been proven to be viable by biological or medical experiments. CONCLUSIONS:Different from traditional methods for judging drug synergy and antagonism, we propose the framework of how to enhance the efficiency by perturbing two sensitive targets in a combinatorial way, how to decrease the drug dose and therefore its side effect and cost by perturbing combinatorially a main sensitive target and an auxiliary insensitive target, and how to perturb two insensitive targets to realize the transition from a disease state to a healthy one which cannot be realized by perturbing each insensitive target alone. Although the idea is mainly applied to an AA metabolic network, the strategy holds for more general molecular networks such as combinatorial regulation in gene regulatory networks.
Project description:MOTIVATION: Drug effects are mainly caused by the interactions between drug molecules and their target proteins including primary targets and off-targets. Identification of the molecular mechanisms behind overall drug-target interactions is crucial in the drug design process. RESULTS: We develop a classifier-based approach to identify chemogenomic features (the underlying associations between drug chemical substructures and protein domains) that are involved in drug-target interaction networks. We propose a novel algorithm for extracting informative chemogenomic features by using L(1) regularized classifiers over the tensor product space of possible drug-target pairs. It is shown that the proposed method can extract a very limited number of chemogenomic features without loosing the performance of predicting drug-target interactions and the extracted features are biologically meaningful. The extracted substructure-domain association network enables us to suggest ligand chemical fragments specific for each protein domain and ligand core substructures important for a wide range of protein families. AVAILABILITY: Softwares are available at the supplemental website. CONTACT: firstname.lastname@example.org SUPPLEMENTARY INFORMATION: Datasets and all results are available at http://cbio.ensmp.fr/~yyamanishi/l1binary/ .
Project description:BackgroundIdentifying drug-target interaction is a key element in drug discovery. In silico prediction of drug-target interaction can speed up the process of identifying unknown interactions between drugs and target proteins. In recent studies, handcrafted features, similarity metrics and machine learning methods have been proposed for predicting drug-target interactions. However, these methods cannot fully learn the underlying relations between drugs and targets. In this paper, we propose anew framework for drug-target interaction prediction that learns latent features from drug-target interaction network.ResultsWe present a framework to utilize the network topology and identify interacting and non-interacting drug-target pairs. We model the problem as a semi-bipartite graph in which we are able to use drug-drug and protein-protein similarity in a drug-protein network. We have then used a graph labeling method for vertex ordering in our graph embedding process. Finally, we employed deep neural network to learn the complex pattern of interacting pairs from embedded graphs. We show our approach is able to learn sophisticated drug-target topological features and outperforms other state-of-the-art approaches.ConclusionsThe proposed learning model on semi-bipartite graph model, can integrate drug-drug and protein-protein similarities which are semantically different than drug-protein information in a drug-target interaction network. We show our model can determine interaction likelihood for each drug-target pair and outperform other heuristics.
Project description:Tuberculosis is a chronic infectious disease caused by Mycobacterium tuberculosis (Mtb). Due to the extensive use of anti-tuberculosis drugs and the development of mutations, the emergence and spread of multidrug-resistant tuberculosis is recognized as one of the most dangerous threats to global tuberculosis control. Some single mutations have been identified to be significantly linked with drug resistance. However, the prior research did not take gene-gene interactions into account, and the emergence of transmissible drug resistance is connected with multiple genetic mutations. In this study we use the bioinformatics software GBOOST (The Hong Kong University, Clear Water Bay, Kowloon, Hong Kong, China) to calculate the interactions of Single Nucleotide Polymorphism (SNP) pairs and identify gene pairs associated with drug resistance. A large part of the non-synonymous mutations in the drug target genes that were included in the screened gene pairs were confirmed by previous reports, which lent sound solid credits to the effectiveness of our method. Notably, most of the identified gene pairs containing drug targets also comprise Pro-Pro-Glu (PPE) family proteins, suggesting that PPE family proteins play important roles in the drug resistance of Mtb. Therefore, this study provides deeper insights into the mechanisms underlying anti-tuberculosis drug resistance, and the present method is useful for exploring the drug resistance mechanisms for other microorganisms.