Abstract: MOTIVATION: Transcriptomes are routinely used to prioritize genes underlying specific phenotypes. Current approaches largely focus on differentially expressed genes (DEGs), despite the recognition that phenotypes emerge via a network of interactions between genes and proteins, many of which may not be differentially expressed. Furthermore, many practical applications lack sufficient samples or an appropriate control to robustly identify statistically significant DEGs. RESULTS: We provide a computational tool-PathExt, which, in contrast to differential genes, identifies differentially active paths when a control is available, and most active paths otherwise, in an omics-integrated biological network. The sub-network comprising such paths, referred to as the TopNet, captures the most relevant genes and processes underlying the specific biological context. The TopNet forms a well-connected graph, reflecting the tight orchestration in biological systems. Two key advantages of PathExt are (i) it can extract characteristic genes and pathways even when only a single sample is available, and (ii) it can be used to study a system even in the absence of an appropriate control. We demonstrate the utility of PathExt via two diverse sets of case studies, to characterize (i) Mycobacterium tuberculosis response upon exposure to 18 antibacterial drugs where only one transcriptomic sample is available for each exposure; and (ii) tissue-relevant genes and processes using transcriptomic data for 39 human tissues. Overall, PathExt is a general tool for prioritizing context-relevant genes in any omics-integrated biological network for any condition(s) of interest, even with a single sample or in the absence of appropriate controls. AVAILABILITYAND IMPLEMENTATION: The source code for PathExt is available at https://github.com/NarmadaSambaturu/PathExt. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Abstract: With the development of high-throughput sequencing technology, biological sequence data reflecting life information becomes increasingly accessible. Particularly on the background of the COVID-19 pandemic, biological sequence data play an important role in detecting diseases, analyzing the mechanism and discovering specific drugs. In recent years, pretraining models that have emerged in natural language processing have attracted widespread attention in many research fields not only to decrease training cost but also to improve performance on downstream tasks. Pretraining models are used for embedding biological sequence and extracting feature from large biological sequence corpus to comprehensively understand the biological sequence data. In this survey, we provide a broad review on pretraining models for biological sequence data. Moreover, we first introduce biological sequences and corresponding datasets, including brief description and accessible link. Subsequently, we systematically summarize popular pretraining models for biological sequences based on four categories: CNN, word2vec, LSTM and Transformer. Then, we present some applications with proposed pretraining models on downstream tasks to explain the role of pretraining models. Next, we provide a novel pretraining scheme for protein sequences and a multitask benchmark for protein pretraining models. Finally, we discuss the challenges and future directions in pretraining models for biological sequences.

Abstract: The use of sodium-glucose co-transporter 2 inhibitors to treat heart failure with preserved ejection fraction (HFpEF) is under investigation in ongoing clinical trials, but the exact mechanism of action is unclear. Here we aimed to use artificial intelligence (AI) to characterize the mechanism of action of empagliflozin in HFpEF at the molecular level. We retrieved information regarding HFpEF pathophysiological motifs and differentially expressed genes/proteins, together with empagliflozin target information and bioflags, from specialized publicly available databases. Artificial neural networks and deep learning AI were used to model the molecular effects of empagliflozin in HFpEF. The model predicted that empagliflozin could reverse 59% of the protein alterations found in HFpEF. The effects of empagliflozin in HFpEF appeared to be predominantly mediated by inhibition of NHE1 (Na(+)/H(+) exchanger 1), with SGLT2 playing a less prominent role. The elucidated molecular mechanism of action had an accuracy of 94%. Empagliflozin's pharmacological action mainly affected cardiomyocyte oxidative stress modulation, and greatly influenced cardiomyocyte stiffness, myocardial extracellular matrix remodelling, heart concentric hypertrophy, and systemic inflammation. Validation of these in silico data was performed in vivo in patients with HFpEF by measuring the declining plasma concentrations of NOS2, the NLPR3 inflammasome, and TGF-beta1 during 12 months of empagliflozin treatment. Using AI modelling, we identified that the main effect of empagliflozin in HFpEF treatment is exerted via NHE1 and is focused on cardiomyocyte oxidative stress modulation. These results support the potential use of empagliflozin in HFpEF.

Abstract: BACKGROUND: Recent efforts in the field of nutritional science have allowed the discovery of disease-beating molecules within foods based on the commonality of bioactive food molecules to FDA-approved drugs. The pioneering work in this field used an unsupervised network propagation algorithm to learn the systemic-wide effect on the human interactome of 1962 FDA-approved drugs and a supervised algorithm to predict anticancer therapeutics using the learned representations. Then, a set of bioactive molecules within foods was fed into the model, which predicted molecules with cancer-beating potential.The employed methodology consisted of disjoint unsupervised feature generation and classification tasks, which can result in sub-optimal learned drug representations with respect to the classification task. Additionally, due to the disjoint nature of the tasks, the employed approach proved cumbersome to optimize, requiring testing of thousands of hyperparameter combinations and significant computational resources.To overcome the technical limitations highlighted above, we represent each drug as a graph (human interactome) with its targets as binary node features on the graph and formulate the problem as a graph classification task. To solve this task, inspired by the success of graph neural networks in graph classification problems, we use an end-to-end graph neural network model operating directly on the graphs, which learns drug representations to optimize model performance in the prediction of anticancer therapeutics. RESULTS: The proposed model outperforms the baseline approach in the anticancer therapeutic prediction task, achieving an F1 score of 67.99%+/-2.52% and an AUPR of 73.91%+/-3.49%. It is also shown that the model is able to capture knowledge of biological pathways to predict anticancer molecules based on the molecules' effects on cancer-related pathways. CONCLUSIONS: We introduce an end-to-end graph convolutional model to predict cancer-beating molecules within food. The introduced model outperforms the existing baseline approach, and shows interpretability, paving the way to the future of a personalized nutritional science approach allowing the development of nutrition strategies for cancer prevention and/or therapeutics.

Abstract: The objective of this study was to determine the intrinsic drug dissolution rate (IDR) and the solute effective transport rate of some drugs, using a single particle dissolution technique, satisfying qualified dissolution conditions. The IDR of three poorly water-soluble compounds was measured in milli-Q water using four different fluid velocities. The enveloped surface area of the particles was calculated from the projected area and the perimeter of the particle observed in the microscope. Furthermore, computational fluid dynamics (CFD) simulations were used to theoretically investigate the flow conditions and dissolution rate, comparing box shaped particles and spherical particles with similar dimensions and surface area as the particles used the experiments. In this study, the IDR measurement of the single particles was determined within 5-60 min using particles with an initial projected area diameter (Dp) between 37.5-104.6 microm. The micropipette-assisted microscopy technique showed a good reproducibility between individual measurements, and the CFD simulations indicated a laminar flow around the particles at all flow velocities, even though there were evident differences in local particle dissolution rates. In conclusion, the IDR and solute effective transport rate were determined under well-defined fluid flow conditions. This type of approach can be used as a complementary approach to traditional dissolution studies to gain in-depth insights into the dissolution process of drug particles.

Abstract: MOTIVATION: Many high-throughput screening studies have been carried out in cancer cell lines to identify therapeutic agents and targets. Existing consistency assessment studies only examined two datasets at a time, with conclusions based on a subset of carefully selected features rather than considering global consistency of all the data. However, poor concordance can still be observed for a large part of the data even when selected features are highly consistent. RESULTS: In this study we assembled nine compound screening datasets and three functional genomics datasets. We derived direct measures of consistency as well as indirect measures of consistency based on association between functional data and copy number-adjusted gene expression data. These results have been integrated into a web application - the Functional Data Consistency Explorer (FDCE), to allow users to make queries and generate interactive visualizations so that functional data consistency can be assessed for individual features of interest. AVAILABILITY: The FDCE web tool and we have developed and the functional data consistency measures we have generated are available at https://lccl.shinyapps.io/FDCE/. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Abstract: Extracts from the plants Phlomis umbrosa and Dipsacus asperoides-which are widely used in Korean and Chinese traditional medicine to treat osteoarthritis and other bone diseases-were used to treat experimental osteoarthritis (OA) rats. Genome-wide differential methylation regions (DMRs) of these medicinal-plant-treated rats were profiled as therapeutic evidence associated with traditional medicine, and they need to be investigated further using detailed molecular research to extrapolate traditional practices to modern medicine. In total, 49 protein-encoding genes whose expression is differentially regulated during disease progression and recovery have been discovered via systematic bioinformatic analysis and have been approved/proposed as druggable targets for various bone diseases by the US food and drug administration. Genes encoding proteins involved in the PI3K/AKT pathway were found to be enriched, likely as this pathway plays a crucial role during OA progression as well as during the recovery process after treatment with the aforementioned plant extracts. The four sub-networks of PI3K/AKT were highly regulated by these plant extracts. Overall, 29 genes were seen in level 2 (51-75%) DMRs and were correlated highly with OA pathogenesis. Here, we propose that these genes could serve as targets to study OA; moreover, the iridoid and triterpenoid phytochemicals obtained from these two plants may serve as potential therapeutic agents.

Abstract: Since neurodevelopmental disorders (NDDs) influence more than 3% of children worldwide, there has been intense investigation to understand the etiology of disorders and develop treatments. Although there are drugs such as aripiprazole, risperidone, and lurasidone, these medications are not cures for the disorders and can only help people feel better or alleviate their symptoms. Thus, it is required to discover therapeutic targets in order to find the ultimate treatments of neurodevelopmental disorders. It is suggested that abnormal neuronal morphology in the neurodevelopment process is a main cause of NDDs, in which the serotonergic system is emerging as playing a crucial role. From this point of view, we noticed the correlation between serotonin receptor subtype 7 (5-HT7R) and NDDs including autism spectrum disorder (ASD), fragile X syndrome (FXS), and Rett syndrome (RTT). 5-HT7R modulators improved altered behaviors in animal models and also affected neuronal morphology via the 5-HT7R/G12 signaling pathway. Through the investigation of recent studies, it is suggested that 5-HT7R could be a potential therapeutic target for the treatment of NDDs.

Abstract: Combinatorial therapies that target multiple pathways have shown great promises for treating complex diseases. DrugComb (https://drugcomb.org/) is a web-based portal for the deposition and analysis of drug combination screening datasets. Since its first release, DrugComb has received continuous updates on the coverage of data resources, as well as on the functionality of the web server to improve the analysis, visualization and interpretation of drug combination screens. Here, we report significant updates of DrugComb, including: (i) manual curation and harmonization of more comprehensive drug combination and monotherapy screening data, not only for cancers but also for other diseases such as malaria and COVID-19; (ii) enhanced algorithms for assessing the sensitivity and synergy of drug combinations; (iii) network modelling tools to visualize the mechanisms of action of drugs or drug combinations for a given cancer sample and (iv) state-of-the-art machine learning models to predict drug combination sensitivity and synergy. These improvements have been provided with more user-friendly graphical interface and faster database infrastructure, which make DrugComb the most comprehensive web-based resources for the study of drug sensitivities for multiple diseases.

Abstract: AIM/HYPOTHESIS: The complexity and heterogeneity of multiple pathological features make Alzheimer's disease (AD) a major culprit to global health. Drug repurposing is an inexpensive and reliable approach to redirect the existing drugs for new indications. The current study aims to study the possibility of repurposing approved anticancer drugs for AD treatment. We proposed an in silico pipeline based on "omics" data mining that combines genomics, transcriptomics, and metabolomics studies. We aimed to validate the neuroprotective properties of repurposed drugs and to identify the possible mechanism of action of the proposed drugs in AD. RESULTS: We generated a list of AD-related genes and then searched DrugBank database and Therapeutic Target Database to find anticancer drugs related to potential AD targets. Specifically, we researched the available approved anticancer drugs and excluded the information of investigational and experimental drugs. We developed a computational pipeline to prioritize the anticancer drugs having a close association with AD targets. From data mining, we generated a list of 2914 AD-related genes and obtained 49 potential druggable targets by functional enrichment analysis. The protein-protein interaction (PPI) studies for these genes revealed 641 interactions. We found that 15 AD risk/direct PPI genes were associated with 30 approved oncology drugs. The computational validation of candidate drug-target interactions, structural and functional analysis, investigation of related molecular mechanisms, and literature-based analysis resulted in four repurposing candidates, of which three drugs were epidermal growth factor receptor (EGFR) inhibitors. CONCLUSION: Our computational drug repurposing approach proposed EGFR inhibitors as potential repurposing drugs for AD. Consequently, our proposed framework could be used for drug repurposing for different indications in an economical and efficient way.