Abstract: Identifying drug-target interaction (DTI) is the basis for drug development. However, the method of using biochemical experiments to discover drug-target interactions has low coverage and high costs. Many computational methods have been developed to predict potential drug-target interactions based on known drug-target interactions, but the accuracy of these methods still needs to be improved. In this article, a graph autoencoder approach for DTI prediction (GADTI) was proposed to discover potential interactions between drugs and targets using a heterogeneous network, which integrates diverse drug-related and target-related datasets. Its encoder consists of two components: a graph convolutional network (GCN) and a random walk with restart (RWR). And the decoder is DistMult, a matrix factorization model, using embedding vectors from encoder to discover potential DTIs. The combination of GCN and RWR can provide nodes with more information through a larger neighborhood, and it can also avoid over-smoothing and computational complexity caused by multi-layer message passing. Based on the 10-fold cross-validation, we conduct three experiments in different scenarios. The results show that GADTI is superior to the baseline methods in both the area under the receiver operator characteristic curve and the area under the precision-recall curve. In addition, based on the latest Drugbank dataset (V5.1.8), the case study shows that 54.8% of new approved DTIs are predicted by GADTI.

Abstract: Millions of deaths caused by Mycobacterium tuberculosis (Mtb) are reported worldwide every year. Treatment of tuberculosis (TB) involves the use of multiple antibiotics over a prolonged period. However, the emergence of resistance leading to multidrug-resistant TB (MDR-TB) and extensively drug-resistant TB (XDR-TB) is the most challenging aspect of TB treatment. Therefore, there is a constant need to search for novel therapeutic strategies that could tackle the growing problem of drug resistance. One such strategy could be perturbing the functions of novel targets in Mtb, such as universal stress protein (USP, Rv1636), which binds to cAMP with a higher affinity than ATP. Orthologs of these proteins are conserved in all mycobacteria and act as "sink" for cAMP, facilitating the availability of this second messenger for signaling when required. Here, we have used the cAMP-bound crystal structure of USP from Mycobacterium smegmatis, a closely related homolog of Mtb, to conduct a structure-guided hunt for potential binders of Rv1636, primarily employing molecular docking approach. A library of 1.9 million compounds was subjected to virtual screening to obtain an initial set of ~2,000 hits. An integrative strategy that uses the available experimental data and consensus indications from other computational analyses has been employed to prioritize 22 potential binders of Rv1636 for experimental validations. Binding affinities of a few compounds among the 22 prioritized compounds were tested through microscale thermophoresis assays, and two compounds of natural origin showed promising binding affinities with Rv1636. We believe that this study provides an important initial guidance to medicinal chemists and biochemists to synthesize and test an enriched set of compounds that have the potential to inhibit Mtb USP (Rv1636), thereby aiding the development of novel antitubercular lead candidates.

Abstract: Drug repositioning is used to find new uses for existing drugs, effectively shortening the drug research and development cycle and reducing costs and risks. A new model of drug repositioning based on ensemble learning is proposed. This work develops a novel computational drug repositioning approach called CMAF to discover potential drug-disease associations. First, for new drugs and diseases or unknown drug-disease pairs, based on their known neighbor information, an association probability can be obtained by implementing the weighted K nearest known neighbors (WKNKN) method and improving the drug-disease association information. Then, a new drug similarity network and new disease similarity network can be constructed. Three prediction models are applied and ensembled to enable the final association of drug-disease pairs based on improved drug-disease association information and the constructed similarity network. The experimental results demonstrate that the developed approach outperforms recent state-of-the-art prediction models. Case studies further confirm the predictive ability of the proposed method. Our proposed method can effectively improve the prediction results.

Abstract: Drug repositioning is an application-based solution based on mining existing drugs to find new targets, quickly discovering new drug-disease associations, and reducing the risk of drug discovery in traditional medicine and biology. Therefore, it is of great significance to design a computational model with high efficiency and accuracy. In this paper, we propose a novel computational method MGRL to predict drug-disease associations based on multi-graph representation learning. More specifically, MGRL first uses the graph convolution network to learn the graph representation of drugs and diseases from their self-attributes. Then, the graph embedding algorithm is used to represent the relationships between drugs and diseases. Finally, the two kinds of graph representation learning features were put into the random forest classifier for training. To the best of our knowledge, this is the first work to construct a multi-graph to extract the characteristics of drugs and diseases to predict drug-disease associations. The experiments show that the MGRL can achieve a higher AUC of 0.8506 based on five-fold cross-validation, which is significantly better than other existing methods. Case study results show the reliability of the proposed method, which is of great significance for practical applications.

Abstract: Protein phosphorylation is one of the best known post-translational mechanisms playing a key role in the regulation of cellular processes. Over 100,000 distinct phosphorylation sites have been discovered through constant improvement of mass spectrometry based phosphoproteomics in the last decade. However, data saturation is occurring and the bottleneck of assigning biologically relevant functionality to phosphosites needs to be addressed. There has been finite success in using data-driven approaches to reveal phosphosite functionality due to a range of limitations. The alternate, more suitable approach is making use of prior knowledge from literature-derived databases. Here, we analysed seven widely used databases to shed light on their suitability to provide functional insights into phosphoproteomics data. We first determined the global coverage of each database at both the protein and phosphosite level. We also determined how consistent each database was in its phosphorylation annotations compared to a global standard. Finally, we looked in detail at the coverage of each database over six experimental datasets. Our analysis highlights the relative strengths and weaknesses of each database, providing a guide in how each can be best used to identify biological mechanisms in phosphoproteomic data.

Abstract: Background: A limited number of targeted therapy options exist for papillary thyroid cancer (PTC) to date. Based on genetic alterations reported by the "The Cancer Genome Atlas (TCGA)", we explored whether PTC shows alterations that may be targetable by drugs approved by the FDA for other solid cancers. Methods: Databases of the National Cancer Institute and MyCancerGenome were screened to identify FDA-approved drugs for targeted therapy. Target genes were identified using Drugbank. Genetic alterations were classified into conferring drug sensitivity or resistance using MyCancerGenome, CiViC, TARGET, and OncoKB. Genomic data for PTC were extracted from TCGA and mined for alterations predicting drug response. Results: A total of 129 FDA-approved drugs with 128 targetable genes were identified. One hundred ninety-six (70%) of 282 classic, 21 (25%) of 84 follicular, and all 30 tall-cell variant PTCs harbored druggable alterations: 259 occurred in 29, 39 in 19, and 31 in 2 targetable genes, respectively. The BRAF V600 mutation was seen in 68% of classic, 16% of follicular variant, and 93% of tall-cell variant PTCs. The RET gene fusion was seen in 8% of classic PTCs, NTRK1 and 3 gene fusions in 3%, and other alterations in <2% of classic variant PTCs. Ninety-nine of 128 (77%) FDA-approved targetable genes did not show any genetic alteration in PTC. Beside selective and non-selective BRAF-inhibitors, no other FDA-approved drug showed any frequent predicted drug sensitivity (<10%). Conclusion: Treatment strategies need to focus on resistance mechanisms to BRAF inhibition and on genetic alteration-independent alternatives rather than on current targeted drugs.

Abstract: p38gamma is a member of the p38 Mitogen Activated Protein Kinases (p38 MAPKs). It contains four subtypes in mammalian cells encoded by different genes including p38alpha (MAPK14), p38beta (MAPK11), p38gamma (MAPK12), and p38delta (MAPK13). Recent studies revealed that p38gamma may exhibit a crucial role in tumorigenesis and cancer aggressiveness. Despite the large number of published literatures, further researches are demanded to clarify its role in cancer development, the tissue-specific function and associated novel treatment strategies. In this article, we provide the latest view on the connection between p38gamma and malignant tumors, highlighting the function of p38gamma. The clinical value of p38gamma is also discussed, helping the translation into the remarkable therapeutic strategy in tumor diseases.

Abstract: Diabetic gastroparesis (DGP) is a serious and chronic complication of long-standing diabetes mellitus, which brings a heavy burden to individuals and society. Traditional Chinese medicine (TCM) is considered a complementary and alternative therapy for DGP patients. Huanglian (Coptidis Rhizoma, HL) and Banxia (Pinelliae Rhizoma, BX) combined as herb pair have been frequently used in TCM prescriptions, which can effectively treat DGP in China. In this article, a practical application of TCM network pharmacological approach was used for the research on herb pair HL-BX in the treatment of DGP. Firstly, twenty-seven potential active components of HL-BX were screened from the TCMSP database, and their potential targets were also retrieved. Then, the compound-target network and PPI network were constructed from predicted common targets, and several key targets were found based on the degree of the network. Next, GO and KEGG enrichment analyses were conducted to obtain several significantly enriched terms. Finally, the experimental verification was made. The results demonstrated that network pharmacological approach was a powerful means for identifying bioactive ingredients and mechanisms of action for TCM. Network pharmacology provided an effective strategy for TCM modern research.

Abstract: Gelsemium elegans (Gardner and Champ.) Benth. (Gelsemiaceae) (GEB) is a toxic plant indigenous to Southeast Asia especially China, and has long been used as Chinese folk medicine for the treatment of various types of pain, including neuropathic pain (NPP). Nevertheless, limited data are available on the understanding of the interactions between ingredients-targets-pathways. The present study integrated network pharmacology and experimental evidence to decipher molecular mechanisms of GEB against NPP. The candidate ingredients of GEB were collected from the published literature and online databases. Potentially active targets of GEB were predicted using the SwissTargetPrediction database. NPP-associated targets were retrieved from GeneCards, Therapeutic Target database, and DrugBank. Then the protein-protein interaction network was constructed. The DAVID database was applied to Gene Ontology and Kyoto Encyclopedia of Genes and Genome pathway enrichment analysis. Molecular docking was employed to validate the interaction between ingredients and targets. Subsequently, a 50 ns molecular dynamics simulation was performed to analyze the conformational stability of the protein-ligand complex. Furthermore, the potential anti-NPP mechanisms of GEB were evaluated in the rat chronic constriction injury model. A total of 47 alkaloids and 52 core targets were successfully identified for GEB in the treatment of NPP. Functional enrichment analysis showed that GEB was mainly involved in phosphorylation reactions and nitric oxide synthesis processes. It also participated in 73 pathways in the pathogenesis of NPP, including the neuroactive ligand-receptor interaction signaling pathway, calcium signaling pathway, and MAPK signaling pathway. Interestingly, 11-Hydroxyrankinidin well matched the active pockets of crucial targets, such as EGFR, JAK1, and AKT1. The 11-hydroxyrankinidin-EGFR complex was stable throughout the entire molecular dynamics simulation. Besides, the expression of EGFR and JAK1 could be regulated by koumine to achieve the anti-NPP action. These findings revealed the complex network relationship of GEB in the "multi-ingredient, multi-target, multi-pathway" mode, and explained the synergistic regulatory effect of each complex ingredient of GEB based on the holistic view of traditional Chinese medicine. The present study would provide a scientific approach and strategy for further studies of GEB in the treatment of NPP in the future.

Abstract: Background: Due to the embryotoxicity found in animal studies and scarce clinical data in pregnant women, it is still controversial whether entecavir (ETV) and adefovir dipivoxil (ADV) are safe during human pregnancy. This is of paramount importance when counseling pregnant women with hepatitis B virus (HBV) on risks and benefits to their offspring. Objective: To quantify the association between administration of ETV and ADV in pregnant women and occurrence of adverse events (AEs) during pregnancy (AEDP). Methods: Pregnancy reports from the FDA Adverse Event Reporting System (FAERS) were used to perform a retrospective analysis of AEDP associated with ETV or ADV. Disproportionality analysis estimating the reporting odds ratio (ROR) was conducted to identify the risk signals. A signal was defined as ROR value >2, and lower limit of 95% confidence interval (CI)> 1. Results: A total of 1,286,367 reports involving AEDP were submitted to FAERS by healthcare professionals. Of these, there were 547 cases reporting ETV and 242 cases reporting ADV as primary suspected drugs. We found a moderate or strong signal for increased risk of spontaneous abortion when comparing ETV with tenofovir disoproxil fumarate (TDF) and telbivudine (LdT), with RORs equal to 1.58 (95% CI, 1.09-2.30) and 2.13 (95% CI, 1.04-4.36), respectively. However, when the included reports were limited to indication containing HBV infection, no signals for increased AEDP were detected. Futhermore, a strong signal for increased risk of spontaneous abortion was identified in patients with HBV infection when comparing ETV or ADV with lamivudine (LAM), with RORs of 3.55 (95% CI, 1.54-8.18) and 2.85 (95% CI, 1.15-7.08), respectively. Conclusion: We found a strong signal for increased risk of spontaneous abortion in patients with HBV infection taking ETV or ADV, in comparison with those prescribed with LAM. Moreover, no obvious signal association of human teratogenicity with exposure to ETV or ADV was identified in fetuses during pregnancy. Nevertheless, owing to the limitations of a spontaneous reporting database, which inevitably contains potential biases, there is a pressing need for well-designed comparative safety studies to validate these results in clinical practice.