A computational study to identify Sesamol derivatives as NRF2 activator for protection against drug-induced liver injury (DILI).

Drug-induced liver injury can be caused by any drugs, their metabolites, or natural products due to the inefficient functioning of drug-metabolizing enzymes, resulting in reactive oxygen species generation and leading to oxidative stress-induced cell death. For protection against oxidative stress, our cell has various defense mechanisms. One of the mechanisms is NRF2 pathway, when activated, protects the cell against oxidative stress. Natural antioxidants such as Sesamol have reported pharmacological activity (hepatoprotective & cardioprotective) and signaling pathways (NRF2 & CREM) altering potential. A Computational analysis was done using molecular docking, IFD, ADMET, MM-GBSA, and Molecular dynamic simulation of the Schrödinger suite. A total of 63,345 Sesamol derivatives were downloaded for the PubChem database. The protein structure of KEAP1-NRF2 (PDB: 4L7D) was downloaded from the RCSB protein database. The molecular docking technique was used to screen compounds that can form an interaction similar to the co-crystalized ligand (1VX). Based on MM-GBSA, docking score, and interactions, ten compounds were selected for ADMET profiling and IFD. After IFD, five compounds (66867225, 46148111, 12444939, 123892179, & 94817569) were selected for molecular dynamics simulation (MDS). Protein-ligand complex stability was assessed during MDS. The selected compounds (66867225, 46148111, 12444939, 123892179, & 94817569) complex with KEAP1 protein shows good stability and bond retentions. In our study, we observed that the selected compounds show good interaction, PCA, Rg, binding free energy, and ADMET profile. We can conclude that the selected compounds can act as NRF2 activators, which should be validated using proper in-vivo/in-vitro models.

Exploring the World of Membrane Proteins: Techniques and Methods for Understanding Structure, Function, and Dynamics.

In eukaryotic cells, membrane proteins play a crucial role. They fall into three categories: intrinsic proteins, extrinsic proteins, and proteins that are essential to the human genome (30% of which is devoted to encoding them). Hydrophobic interactions inside the membrane serve to stabilize integral proteins, which span the lipid bilayer. This review investigates a number of computational and experimental methods used to study membrane proteins. It encompasses a variety of technologies, including electrophoresis, X-ray crystallography, cryogenic electron microscopy (cryo-EM), nuclear magnetic resonance spectroscopy (NMR), biophysical methods, computational methods, and artificial intelligence. The link between structure and function of membrane proteins has been better understood thanks to these approaches, which also hold great promise for future study in the field. The significance of fusing artificial intelligence with experimental data to improve our comprehension of membrane protein biology is also covered in this paper. This effort aims to shed light on the complexity of membrane protein biology by investigating a variety of experimental and computational methods. Overall, the goal of this review is to emphasize how crucial it is to understand the functions of membrane proteins in eukaryotic cells. It gives a general review of the numerous methods used to look into these crucial elements and highlights the demand for multidisciplinary approaches to advance our understanding.

Repositioning of Quinazolinedione-Based Compounds on Soluble Epoxide Hydrolase (sEH) through 3D Structure-Based Pharmacophore Model-Driven Investigation.

The development of new bioactive compounds represents one of the main purposes of the drug discovery process. Various tools can be employed to identify new drug candidates against pharmacologically relevant biological targets, and the search for new approaches and methodologies often represents a critical issue. In this context, in silico drug repositioning procedures are required even more in order to re-evaluate compounds that already showed poor biological results against a specific biological target. 3D structure-based pharmacophoric models, usually built for specific targets to accelerate the identification of new promising compounds, can be employed for drug repositioning campaigns as well. In this work, an in-house library of 190 synthesized compounds was re-evaluated using a 3D structure-based pharmacophoric model developed on soluble epoxide hydrolase (sEH). Among the analyzed compounds, a small set of quinazolinedione-based molecules, originally selected from a virtual combinatorial library and showing poor results when preliminarily investigated against heat shock protein 90 (Hsp90), was successfully repositioned against sEH, accounting the related built 3D structure-based pharmacophoric model. The promising results here obtained highlight the reliability of this computational workflow for accelerating the drug discovery/repositioning processes.

Design and synthesis of pyrazolone-based compounds as potent blockers of SARS-CoV-2 viral entry into the host cells.

SARS-CoV-2 are enveloped positive-stranded RNA viruses that replicate in the cytoplasm. It relies on the fusion of their envelope with the host cell membrane to deliver their nucleocapsid into the host cell. The spike glycoprotein (S) mediates virus entry into cells via the human Angiotensin-converting enzyme 2 (hACE2) protein located on many cell types and tissues' outer surface. This study, therefore, aimed to design and synthesize novel pyrazolone-based compounds as potential inhibitors that would interrupt the interaction between the viral spike protein and the host cell receptor to prevent SARS-CoV 2 entrance into the cell. A series of pyrazolone compounds as potential SARS-CoV-2 inhibitors were designed and synthesized. Employing computational techniques, the inhibitory potentials of the designed compounds against both spike protein and hACE2 were evaluated. Results of the binding free energy from the in-silico analysis, showed that three compounds (7i, 7k and 8f) and six compounds (7b, 7h, 7k, 8d, 8g, and 8h) showed higher and better binding high affinity to SARS-CoV-2 Sgp and hACE-2, respectively compared to the standard drugs cefoperazone (CFZ) and MLN-4760. Furthermore, the outcome of the structural analysis of the two proteins upon binding of the inhibitors showed that the two proteins (SARS-CoV-2 Sgp and hACE-2) were stable, and the structural integrity of the proteins was not compromised. This study suggests pyrazolone-based compounds might be potent blockers of the viral entry into the host cells.

Gradient Profile Estimation Using Exponential Cubic Spline Smoothing in a Bayesian Framework.

Attaining reliable gradient profiles is of utmost relevance for many physical systems. In many situations, the estimation of the gradient is inaccurate due to noise. It is common practice to first estimate the underlying system and then compute the gradient profile by taking the subsequent analytic derivative of the estimated system. The underlying system is often estimated by fitting or smoothing the data using other techniques. Taking the subsequent analytic derivative of an estimated function can be ill-posed. This becomes worse as the noise in the system increases. As a result, the uncertainty generated in the gradient estimate increases. In this paper, a theoretical framework for a method to estimate the gradient profile of discrete noisy data is presented. The method was developed within a Bayesian framework. Comprehensive numerical experiments were conducted on synthetic data at different levels of noise. The accuracy of the proposed method was quantified. Our findings suggest that the proposed gradient profile estimation method outperforms the state-of-the-art methods.

Evaluation of computational techniques for predicting non-synonymous single nucleotide variants pathogenicity.

The human genetic diseases associated with many factors, one of these factors is the non-synonymous Single Nucleotide Variants (nsSNVs) cause single amino acid change with another resulting in protein function change leading to disease. Many computational techniques have been released to expect the impacts of amino acid alteration on protein function and classify mutations as pathogenic or neutral. Here in this article, we assessed the performance of eight techniques; FATHMM, SIFT, Provean, iFish, Mutation Assessor, PANTHER, SNAP2, and PON- P2 using a VaribenchSelectedPure dataset of 2144 pathogenic variants and 3777 neutral variants extracted from the free standard database "Varibench." The first five techniques achieve (45.60-83.75) % specificity, (52.64-94.13) % sensitivity, (51.00-88.90) % AUC, and (49.76-88.24) % ACC on whole dataset, while all eight techniques achieve (36.54-77.88) % specificity, (50.00-75.00) % sensitivity, (51.00-76.40) % AUC, and (25.00-77.78) % ACC on random sample dataset. We also created a Meta classifier (CSTJ48) that combines FATHMM, iFish, and Mutation Assessor. It registers 96.33% specificity, 86.07% sensitivity, 91.20% AUC, and 91.89 ACC. By comparing the results, it's clear that FATHMM gives the highest performance over the seven individual techniques, where it achieves 83.75% and 77.88% specificity, 94.13%, and 75.00% sensitivity, 88.90% and 76.40% AUC, and 88.24% and 77.78% ACC on whole and random sample dataset, respectively. Also, the launched Meta classifier (CSTJ48) is outperforming over all the eight individual tools that compared here.

Current Projection Methods-Induced Biases at Subgroup Detection for Machine-Learning Based Data-Analysis of Biomedical Data.

Advances in flow cytometry enable the acquisition of large and high-dimensional data sets per patient. Novel computational techniques allow the visualization of structures in these data and, finally, the identification of relevant subgroups. Correct data visualizations and projections from the high-dimensional space to the visualization plane require the correct representation of the structures in the data. This work shows that frequently used techniques are unreliable in this respect. One of the most important methods for data projection in this area is the t-distributed stochastic neighbor embedding (t-SNE). We analyzed its performance on artificial and real biomedical data sets. t-SNE introduced a cluster structure for homogeneously distributed data that did not contain any subgroup structure. In other data sets, t-SNE occasionally suggested the wrong number of subgroups or projected data points belonging to different subgroups, as if belonging to the same subgroup. As an alternative approach, emergent self-organizing maps (ESOM) were used in combination with U-matrix methods. This approach allowed the correct identification of homogeneous data while in sets containing distance or density-based subgroups structures; the number of subgroups and data point assignments were correctly displayed. The results highlight possible pitfalls in the use of a currently widely applied algorithmic technique for the detection of subgroups in high dimensional cytometric data and suggest a robust alternative.

Slow Magnetic Relaxation in a Palladium-Gadolinium Complex Induced by Electron Density Donation from the Palladium Ion.

Incorporating palladium in the first coordination sphere of acetato-bridged lanthanoid complexes, [Pd2 Ln2 (H2 O)2 (AcO)10 ]⋅2 AcOH (Ln=Gd (1), Y (2), Gd0.4 Y1.6 (3), Eu (4)), led to significant bonding interactions between the palladium and the lanthanoid ions, which were demonstrated by experimental and theoretical methods. We found that electron density was donated from the d8 Pd2+ ion to Gd3+ ion in 1 and 3, leading to the observed slow magnetic relaxation by using local orbital locator (LOL) and X-ray absorption near-edge structure (XANES) analysis. Field-induced dual slow magnetic relaxation was observed for 1 up to 20 K. Complex 3 and frozen aqueous and acetonitrile solutions of 1 showed only one relaxation peak, which confirms the role of intermolecular dipolar interactions in slowing the magnetic relaxation of 1. The slow magnetic relaxation occurred through a combination of Orbach and Direct processes with the highest pre-exponential factor (τo =0.06 s) reported so far for a gadolinium complex exhibiting slow magnetic relaxation. The results revealed that transition metal-lanthanoid (TM-Ln) axial interactions indeed could lead to new physical properties by affecting both the electronic and magnetic states of the compounds.

In pursuit of an accurate spatial and temporal model of biomolecules at the atomistic level: a perspective on computer simulation.

Despite huge advances in the computational techniques available for simulating biomolecules at the quantum-mechanical, atomistic and coarse-grained levels, there is still a widespread perception amongst the experimental community that these calculations are highly specialist and are not generally applicable by researchers outside the theoretical community. In this article, the successes and limitations of biomolecular simulation and the further developments that are likely in the near future are discussed. A brief overview is also provided of the experimental biophysical methods that are commonly used to probe biomolecular structure and dynamics, and the accuracy of the information that can be obtained from each is compared with that from modelling. It is concluded that progress towards an accurate spatial and temporal model of biomacromolecules requires a combination of all of these biophysical techniques, both experimental and computational.

In silico Approaches for Exploring the Pharmacological Activities of Benzimidazole Derivatives: A Comprehensive Review.

BACKGROUND: This article reviews computational research on benzimidazole derivatives. Cytotoxicity for all compounds against cancer cell lines was measured and the results revealed that many compounds exhibited high inhibitions. This research examines the varied pharmacological properties like anticancer, antibacterial, antioxidant, anti-inflammatory and anticonvulsant activities of benzimidazole derivatives. The suggested method summarises In silico research for each activity. This review examines benzimidazole derivative structure-activity relationships and pharmacological effects. In silico investigations can anticipate structural alterations and their effects on these derivative's pharmacological characteristics and efficacy through many computational methods. Molecular docking, molecular dynamics simulations and virtual screening help anticipate pharmacological effects and optimize chemical design. These trials will improve lead optimization, target selection, and ADMET property prediction in drug development. In silico benzimidazole derivative studies will be assessed for gaps and future research. Prospective studies might include empirical verification, pharmacodynamic analysis, and computational methodology improvement. OBJECTIVES: This review discusses benzimidazole derivative In silico research to understand their specific pharmacological effects. This will help scientists design new drugs and guide future research. METHODS: Latest, authentic and published reports on various benzimidazole derivatives and their activities are being thoroughly studied and analyzed. RESULT: The overview of benzimidazole derivatives is more comprehensive, highlighting their structural diversity, synthetic strategies, mechanisms of action, and the computational tools used to study them. CONCLUSION: In silico studies help to understand the structure-activity relationship (SAR) of benzimidazole derivatives. Through meticulous alterations of substituents, ring modifications, and linker groups, this study identified the structural factors influencing the pharmacological activity of benzimidazole derivatives. These findings enable the rational design and optimization of more potent and selective compounds.

Molecular simulation-based insights into dye pollutant adsorption: A perspective review.

Growing concerns about environmental pollution have highlighted the need for efficient and sustainable methods to remove dye contamination from various ecosystems. In this context, computational methods such as molecular dynamics (MD), Monte Carlo (MC) simulations, quantum mechanics (QM) calculations, and machine learning (ML) methods are powerful tools used to study and predict the adsorption processes of dyes on various adsorbents. These methods provide detailed insights into the molecular interactions and mechanisms involved, which can be crucial for designing efficient adsorption systems. MD simulations, detailing molecular arrangements, predict dyes' adsorption behaviour and interaction energies with adsorbents. They simulate the entire adsorption process, including surface diffusion, solvent layer penetration, and physisorption. QM calculations, especially density functional theory (DFT), determine molecular structures and reactivity descriptors, aiding in understanding adsorption mechanisms. They identify stable adsorption configurations and interactions like hydrogen bonding and electrostatic forces. MC simulations predict equilibrium properties and adsorption energies by sampling molecular configurations. ML methods have proven highly effective in predicting and optimizing dye adsorption processes. These models offer significant advantages over traditional methods, including higher accuracy and the ability to handle complex datasets. These methods optimize adsorption conditions, clarify adsorbent functionalization roles, and predict dye removal efficiency under various conditions. This research explores MD, MC, QM, and ML approaches to connect molecular interactions with macroscopic adsorption phenomena. Probing these techniques provides insights into the dynamics and energetics of dye pollutants on adsorption surfaces. The findings will aid in developing and optimizing new materials for dye removal. This review has significant implications for environmental remediation, offering a comprehensive understanding of adsorption at various scales. Merging microscopic data with macroscopic observations enhances knowledge of dye pollutant adsorption, laying the groundwork for efficient, sustainable removal technologies. Addressing the growing challenges of ecosystem protection, this study contributes to a cleaner, more sustainable future.

Remapping the Chemical Space and the Pharmacological Space of Drugs: What Can We Expect from the Road Ahead?

This work examines the current landscape of drug discovery and development, with a particular focus on the chemical and pharmacological spaces. It emphasizes the importance of understanding these spaces to anticipate future trends in drug discovery. The use of cheminformatics and data analysis enabled in silico exploration of these spaces, allowing a perspective of drugs, approved drugs after 2020, and clinical candidates, which were extracted from the newly released ChEMBL34 (March 2024). This perspective on chemical and pharmacological spaces enables the identification of trends and areas to be occupied, thereby creating opportunities for more effective and targeted drug discovery and development strategies in the future.

Sounding out the dynamics: a concise review of high-speed photoacoustic microscopy.

Significance: Photoacoustic microscopy (PAM) offers advantages in high-resolution and high-contrast imaging of biomedical chromophores. The speed of imaging is critical for leveraging these benefits in both preclinical and clinical settings. Ongoing technological innovations have substantially boosted PAM's imaging speed, enabling real-time monitoring of dynamic biological processes. Aim: This concise review synthesizes historical context and current advancements in high-speed PAM, with an emphasis on developments enabled by ultrafast lasers, scanning mechanisms, and advanced imaging processing methods. Approach: We examine cutting-edge innovations across multiple facets of PAM, including light sources, scanning and detection systems, and computational techniques and explore their representative applications in biomedical research. Results: This work delineates the challenges that persist in achieving optimal high-speed PAM performance and forecasts its prospective impact on biomedical imaging. Conclusions: Recognizing the current limitations, breaking through the drawbacks, and adopting the optimal combination of each technology will lead to the realization of ultimate high-speed PAM for both fundamental research and clinical translation.

Comprehensive views toward the biomolecular recognition of an anticancer drug, leflunomide with human serum albumin.

Biomolecular association of an anticancer drug, leflunomide (LEF) with human serum albumin (HSA), the leading ligands carrier in human circulation was characterized using biophysical (i.e., fluorescence, absorption and voltammetric) methods and computational (i.e., molecular docking and molecular dynamics simulation) techniques. Evaluations of fluorescence, absorption and voltammetric findings endorsed the complex formation between LEF and HSA. An inverse relationship of Stern-Volmer constant-temperature and hyperchromic shift of the protein's absorption signal with addition of LEF confirmed the LEF quenched the HSA fluorescence through static process. Moderate nature of binding strength (binding constant = 2.76-4.77 × 104 M-1) was detected towards the LEF-HSA complexation, while the association process was naturally driven via hydrophobic interactions, van der Waals interactions and hydrogen bonds, as evident from changes in entropy (ΔS= + 19.91 J mol-1 K-1) and enthalpy (ΔH = - 20.09 kJ mol-1), and molecular docking assessments. Spectral analyses of synchronous and three-dimensional fluorescence validated microenvironmental fluctuations near Trp and Tyr residues upon LEF binding to the protein. LEF association with HSA significantly defended temperature-induced destabilization of the protein. Although LEF was found to attach to HSA at Sudlow's sites I and II, but exhibited greater preference toward its site I, as detected by the investigations of competitive site-marker displacement. Molecular dynamics simulation assessment revealed that the complex attained equilibrium throughout simulations, showing the LEF-HSA complex constancy.Communicated by Ramaswamy H. Sarma.

A Comparative Analytical Review on Machine Learning Methods in Drugtarget Interactions Prediction.

BACKGROUND: Predicting drug-target interactions (DTIs) is an important topic of study in the field of drug discovery and development. Since DTI prediction in vitro studies is very expensive and time-consuming, computational techniques for predicting drug-target interactions have been introduced successfully to solve these problems and have received extensive attention. OBJECTIVE: In this paper, we provided a summary of databases that are useful in DTI prediction and intend to concentrate on machine learning methods as a chemogenomic approach in drug discovery. Unlike previous surveys, we propose a comparative analytical framework based on the evaluation criteria. METHODS: In our suggested framework, there are three stages to follow: First, we present a comprehensive categorization of machine learning-based techniques as a chemogenomic approach for drug-target interaction prediction problems; Second, to evaluate the proposed classification, several general criteria are provided; Third, unlike other surveys, according to the evaluation criteria introduced in the previous stage, a comparative analytical evaluation is performed for each approach. RESULTS: This systematic research covers the earliest, most recent, and outstanding techniques in the DTI prediction problem and identifies the advantages and weaknesses of each approach separately. Additionally, it can be helpful in the effective selection and improvement of DTI prediction techniques, which is the main superiority of the proposed framework. CONCLUSION: This paper gives a thorough overview to serve as a guide and reference for other researchers by providing an analytical framework which can help to select, compare, and improve DTI prediction methods.

Computational Analysis of CD46 Protein Interaction with SARS-CoV-2 Structural Proteins: Elucidating a Putative Viral Entry Mechanism into Human Cells.

This study examines an unexplored aspect of SARS-CoV-2 entry into host cells, which is widely understood to occur via the viral spike (S) protein's interaction with human ACE2-associated proteins. While vaccines and inhibitors targeting this mechanism are in use, they may not offer complete protection against reinfection. Hence, we investigate putative receptors and their cofactors. Specifically, we propose CD46, a human membrane cofactor protein, as a potential putative receptor and explore its role in cellular invasion, acting possibly as a cofactor with other viral structural proteins. Employing computational techniques, we created full-size 3D models of human CD46 and four key SARS-CoV-2 structural proteins-EP, MP, NP, and SP. We further developed 3D models of CD46 complexes interacting with these proteins. The primary aim is to pinpoint the likely interaction domains between CD46 and these structural proteins to facilitate the identification of molecules that can block these interactions, thus offering a foundation for novel pharmacological treatments for SARS-CoV-2 infection.

Computational studies of Piezo1 yield insights into key lipid-protein interactions, channel activation, and agonist binding.

Piezo1 is a mechanically gated ion channel responsible for converting mechanical stimuli into electrical signals in mammals, playing critical roles in vascular development and blood pressure regulation. Dysfunction of Piezo1 has been linked to several disorders, including hereditary xerocytosis (gain-of-function) and generalised lymphatic dysplasia (loss-of-function), as well as a common polymorphism associated with protection against severe malaria. Despite the important physiological roles played by Piezo1, its recent discovery means that many aspects underlying its function are areas of active research. The recently elucidated cryo-EM structures of Piezo1 have paved the way for computational studies, specifically molecular dynamic simulations, to examine the protein's behaviour at an atomistic level. These studies provide valuable insights to Piezo1's interactions with surrounding membrane lipids, a small-molecule agonist named Yoda1, and Piezo1's activation mechanisms. In this review, we summarise and discuss recent papers which use computational techniques in combination with experimental approaches such as electrophysiology/mutagenesis studies to investigate Piezo1. We also discuss how to mitigate some shortcomings associated with using computational techniques to study Piezo1 and outline potential avenues of future research.

Proteomic and computational characterisation of 11S globulins from grape seed flour by-product and its interaction with malvidin 3-glucoside by molecular docking.

Grape seed flour by-product (GSBP) is an economic and renewable source of proteins, increasingly being explored due to interesting technological application such as colour protection in rich-anthocyanins beverages. Globulin-like proteins from GSBP were characterised by proteomic and computational studies. MALDI TOF/TOF analysis revealed the presence of two 11S globulins (acid and basic), whose 3D structures have been elucidated for the first time in Vitis vinifera L. grape seeds by using homology models and molecular dynamics. The secondary structure showed 11 α-helices and 25 ß-sheets for acid and 12 α-helices and 24 ß-sheets for basic 11S globulins. Molecular docking results indicate that both grape seed 11S globulins could establish different types of non-covalent interactions (π-π) with malvidin 3-O-glucoside (wine anthocyanin), which suggest a possible colour protection similar to that occurring in copigmentation phenomenon. These findings provide valuable information of globulin family proteins that could be relevant in food industry applications.

Inhibitory mechanism of Ambroxol and Bromhexine Hydrochlorides as potent blockers of molecular interaction between SARS-CoV-2 spike protein and human angiotensin-converting Enzyme-2.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infects the host cells through interaction of its spike protein with human angiotensin-converting enzyme 2 (hACE-2). High binding affinity between the viral spike protein and host cells hACE-2 receptor has been reported to enhance the viral infection. Thus, the disruption of this molecular interaction will lead to reduction in viral infectivity. This study, therefore, aimed to analyze the inhibitory potentials of two mucolytic drugs; Ambroxol hydrochlorides (AMB) and Bromhexine hydrochlorides (BHH), to serve as potent blockers of these molecular interactions and alters the binding affinity/efficiency between the proteins employing computational techniques. The study examined the effects of binding of each drug at the receptor binding domain (RBD) of the spike protein and the exopeptidase site of hACE-2 on the binding affinity (ΔGbind) and molecular interactions between the two proteins. Binding affinity revealed that the binding of the two drugs at the RBD-ACE-2 site does not alter the binding affinity and molecular interaction between the proteins. However, the binding of AMB (-56.931 kcal/mol) and BHH (-46.354 kcal/mol) at the exopeptidase site of hACE-2, significantly reduced the binding affinities between the proteins compared to the unbound, ACE-2-RBD complex (-64.856 kcal/mol). The result further showed the two compounds have good affinity at the hACE-2 site, inferring they might be potent inhibitors of hACE-2. Residue interaction networks analysis further revealed the binding of the two drugs at the exopeptidase site of hACE-2 reduced the number of interacting amino residues, subsequently leading to loss of interactions between the two proteins, with BHH showing better reduction in the molecular interaction and binding affinity than AMB. The result of the structural analyses additionally, revealed that the binding of the drugs considerably influences the dynamic of the complexes when compared to the unbound complex. The findings from this study suggest the binding of the two drugs at the exopeptidase site reduces the binding effectiveness of the proteins than their binding at the RBD site, and consequently might inhibit viral attachment and entry.

Predicting non-small cell lung cancer-related genes by a new network-based machine learning method.

Lung cancer is the leading cause of cancer death globally, killing 1.8 million people yearly. Over 85% of lung cancer cases are non-small cell lung cancer (NSCLC). Lung cancer running in families has shown that some genes are linked to lung cancer. Genes associated with NSCLC have been found by next-generation sequencing (NGS) and genome-wide association studies (GWAS). Many papers, however, neglected the complex information about interactions between gene pairs. Along with its high cost, GWAS analysis has an obvious drawback of false-positive results. Based on the above problem, computational techniques are used to offer researchers alternative and complementary low-cost disease-gene association findings. To help find NSCLC-related genes, we proposed a new network-based machine learning method, named deepRW, to predict genes linked to NSCLC. We first constructed a gene interaction network consisting of genes that are related and irrelevant to NSCLC disease and used deep walk and graph convolutional network (GCN) method to learn gene-disease interactions. Finally, deep neural network (DNN) was utilized as the prediction module to decide which genes are related to NSCLC. To evaluate the performance of deepRW, we ran tests with 10-fold cross-validation. The experimental results showed that our method greatly exceeded the existing methods. In addition, the effectiveness of each module in deepRW was demonstrated in comparative experiments.