Journal of Computational Biophysics and Chemistry

Metastasis remains the reason for chemoresistance and high cancer mortality. It is hence a valuable predictive factor in cancer prognosis and drug sensitivity. Single-cell RNA sequencing (scRNA-seq) can reveal cellular heterogeneity in metastasis microenvironment and capture high-resolution signatures for improved cancer prediction. As a case study, an integrated analysis framework was designed for metastatic lung adenocarcinoma (LUAD) scRNA-seq profiles and we identified nine key prognostic genes (KPGs) that were trained and validated in 407 internal and external patient cohorts using machine learning and other methods. Correlation analysis revealed the strong association between KPGs signatures and several clinical characteristics such as gender, T-stage and N-stage. We incorporated these risk clinical variables into a KPGs nomogram model with superior accuracy for overall survival (OS) prediction. We also found that high risk group with high nomogram scores had poorer prognosis accompanied by a higher tumor mutation burden (TMB) and was more sensitive to chemotherapy and targeted agents, which was associated with the upregulation of DNA replication, ECM receptor interaction, P53 signaling pathway, spliceosome and proteasome pathway. Collectively, we proposed a simple and feasible strategy to mine single-cell resolution metastasis signatures from scRNA-seq data for improved cancer prognosis and drug sensitivity prediction, which will be a useful tool in risk gene discovery and targeted therapy in metastatic cancers.

Alzheimer’s disease is a highly prevalent disease that has affected over 50 million people worldwide. It is mainly caused by the continuous deposition of amyloid-$\beta$ peptide (A$\beta )$. Acetylcholinesterase (AChE), N-methyl D-aspartate (NMDA), Tau-Tubulin kinase 1 (TTBK1) and $\beta$-secretase 1 (BACE-1) are potential targets to treat Alzheimer’s disease and to reduce the pathogenesis; amongst, BACE-1 is preferred owing to its capacity in minimizing the accumulation of A$\beta$. This study explored Semecarpus anacardium Linn.-derived phytocompounds and their interaction with BACE-1. The phytochemicals of S. anacardium Linn. were retrieved from the IMPPAT database. The factual targets (AChE, BACE-1, NMDA and TTBK1) of Alzheimer’s disease are identified by the graph theoretical network analysis, and are docked with a total of eight bioactive compounds from S. anacardium Linn. Interestingly, the docking score was significant (−12 kcal × mol${}^{-1})$ for the interactions of amentoflavone with BACE-1. The co-crystallized ligand of BACE-1 (2-imino-3-methyl-5,5-diphenylimidazolidin-4-one) was considered for further studies and revealed a superior binding affinity (−8.7 kcal × mol${}^{-1})$. Molecular dynamics (MD) simulation and Density Functional Theory (DFT) analyzed the conformational stability of amentoflavone and BACE-1 complex. Furthermore, S. anacardium Linn. seemed to be nontoxic against U87-MG cells in vitro. These results from in silico prediction suggest that the derivatives of S. anacardium Linn. possess a significant interaction with the BACE-1 and provide a favorable direction to prevent aggregation of A$\beta$ peptides.

Epilepsy, one of the most common neurological disorders, is characterized by continuing predisposition of epileptic seizures. Abnormal discharges of seizures burst the electrical functions in the brain leading to unprompted behaviors of brain activity. Recently, more attention has been focused on bioactive substances from plants as therapeutic agents for neurodegenerative disorders because of their efficacy, low toxicity, and less side-effects. This study has utilized in silico molecular modeling as a tool for screening effective antiepileptic compounds from Sida cordifolia which could be further explored for the treatment of epilepsy. Twenty-six bioactive compounds were identified through data repository and their structures were retrieved from PubChem database. Based on the Graph theoretical network analysis, NMDA protein was identified as an ideal drug target. The effective compound was chosen based on the docking scores and its drug-likeness properties were analyzed by in silico prediction of pharmacokinetic and physicochemical properties. Based on molecular docking, 5,7-dihydroxy-3-isoprenyl flavone was identified as a better binder as this exhibited higher binding towards the target with lower binding energy (−10.5 kcal × mol${}^{-1})$. The stability of the protein–ligand complex was confirmed by molecular dynamics simulation study. The intermolecular interaction assessed during dynamic conditions indicated that the bioactive compound, 5,7-dihydroxy-3-isoprenyl flavone, from Sida cordifolia could be a potential lead molecule that can be developed as a candidate drug for anti-epileptic therapy.

The global spread of COVID-19 caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) originated in Wuhan in December 2019, created a massive health crisis, and disrupted the world economy. Much research has been conducted to discover drugs, develop vaccines, and find repurposable drugs against the disease. Computational drug repurposing, the process of determining new uses for approved drugs through computational techniques, becomes an effective solution to fight the COVID-19 pandemic. This study aims to investigate and prioritize potential drugs against SARS-CoV-2 through an integrated network-based approach. We propose an ensemble approach based on network inference and inductive matrix completion (NIMCVDA) for virus–drug association prediction to identify antivirals against COVID-19. We constructed a heterogeneous drug–virus network using intra-similarities of virus genomic sequences and drug chemical structures and existing associations between viruses and drugs. A network inference method is used to infer missing drug–virus edges. Based on this, existing drug–virus association matrix is reconstructed. Finally, more accurate association scores between drugs and viruses are computed using the inductive matrix completion algorithm. The proposed method achieved an AUC of 0.9020 on five-fold cross-validation and 0.8786 on leave-one-out cross-validation. We compared the performance of the model with related approaches. In addition, we carried out case studies on the top-predicted drugs and implemented our model with other datasets to verify prediction performance. Our work prioritized repurposable drugs to battle with COVID-19 epidemic. The cross-validation results and case studies illustrate that the top-predicted drugs are strong candidates for further biological tests.

Traditional drug development requires a lot of time and effort, while computational drug repositioning enables to discover the underlying mechanisms of drugs, thereby reducing the cost of drug discovery and development. However, existing computational models utilize only low-order biological information at the level of individual drugs, diseases and their associations, making it difficult to dig deeper and fuse higher-order structural information. In this paper, we develop a novel prediction model, namely DRGDL, based on graph deep learning to infer potential drug-disease associations (DDAs) by coming chemical structural similarity information, which can more comprehensively make use of the biological features of drugs and diseases. First, the lower-order and higher-order representations of drugs and diseases are captured by two different graph deep learning strategies. Then, suitable negative samples are select by use of k-means algorithm. At last, the Random Forest classifier is applied to complete the prediction task of DDAs by integrating two representations of drugs and diseases. Experiment results indicate that DRGDL achieves excellent performance under ten-fold cross-validation on two benchmark datasets.

Mental health counseling is a significant challenge in contemporary society, primarily due to issues such as cost, stigma, fear, and limited availability. Emotions play a crucial role in conveying information in this context, making emotion detection essential for a deeper understanding of an individual’s mental well-being. Utilizing generative machine learning models in mental health counseling could potentially lower barriers to access and improve outcomes. This paper proposes the development of a deep learning-based emotion-detecting chatbot named Serenity. The approach involves combining a pre-trained deep neural model, RoBERTa, with a multi-resolution adversarial model, EmpDG, to enhance the accuracy of detected emotions and generate more empathetic responses. RoBERTa has been trained on a dataset of thousands of tweets from Twitter. Additionally, an interactive adversarial learning framework is introduced to leverage user feedback and assess the emotional perceptivity of generated responses in dialogues. The study aims to demonstrate that a machine learning-based mental health chatbot like Serenity has the potential to serve as an effective complement to traditional human counselors.

Background: Neurofibromatosis Type 1 (NF1) is an autosomal dominant benign tumor predisposition syndrome. Till date very little is known about the pathogenesis and underlying mechanisms of NF1. Objective: The aim of this study is to use integrated bioinformatics approaches to explore the pathways involved in NF1 onset by identifying hub genes and potential drugs. Methods: Microarray profiling datasets were used to obtain Differentially Expressed Genes (DEGs) by analyzing in GEO2R. 1730 DEGs were found and STRING analysis revealed Protein-Protein Interaction (PPI) network of these DEGs. Molecular Complex was detected using MCODE plugin of Cytoscape. CytoHubba plugin of Cytoscape was used to declare hub genes based on their degree of connectivity. Lastly, Drug-Gene Interaction Database (DGIdb) and Gene Expression Profiling Interactive Analysis (GEPIA) servers were used to find potential drugs and survival analysis of hub gene, respectively. Results: Degree of connectivity revealed 10 hub genes out of 1730 DEGs. The survival analysis of these hub genes resulted in three specific genes that are altered in NF1 patients. These three genes are named as BRCA1, KIF11 and LAMA1. Furthermore, 13 potential drugs were also predicted to influence BRCA1, KIF11 and LAMA1 and have therapeutic potential against NF1. Conclusion: This research opens ways to understanding the pathogenesis of NF1. All three genes are key biological markers and potential drug targets that can be confirmed by further experiments. Moreover, 13 potential drugs were found to work against the disease.

In the continuous search for potential drug-like molecules for diseases, drug repurposing offers a solution by identifying new roles for the existing drugs. For breast cancer, which is the leading cause of death among women, discovering a new drug with fewer side effects can take more than a year, and drug repurposing provides a faster alternative. In this study, we used machine learning (ML) and structure-based screening to evaluate molecules from the ChEMBL library against Human Aromatase, an enzyme that is involved in the conversion of androgens to estrogen. The virtual screening output identified 148 potential compounds. These compounds were further subjected to detailed structure-based virtual screening, followed by assessment of induced fit docking, quantum-based docking and ADMET properties. The top compound, CHEMBL502014 and the co-crystallized ligand (reference) were evaluated through MD simulations for 100 ns. The results indicated that the CHEMBL502014-Aromatase complex demonstrated better conformational stability throughout the simulation compared to the reference ligand. Based on these analyses, we propose CHEMBL502014 and other top compounds with known activities against various proteins as potential candidates for anti-cancer activity against breast cancer. These compounds should be further evaluated using experimental assays to confirm their effectiveness against breast cancer.