Computer-Based Solutions to Investigate Biological- and Health-Related Problems

Nonalcoholic fatty liver disease (NAFLD) is the most common chronic liver disease with multifactorial pathogenesis; histone demethylases (HDMs) are emerging as attractive targets. We identified HDM genes (including KDM5C, KDM6B, KDM8, KDM4A, and JMJD7) that were differentially expressed in NAFLD and normal samples by exploring gene expression profiling datasets. There was no significant difference in the expression of genes related to histone demethylation between mild and advanced NAFLD. In vitro and in vivo studies indicated that KDM6B and JMJD7 were upregulated at the mRNA level in NAFLD. We explored the expression levels and prognostic values of the identified HDM genes in hepatocellular carcinoma (HCC). KDM5C and KDM4A were upregulated in HCC compared to normal tissue, while KDM8 showed downregulation. The abnormal expression levels of these HDMs could provide prognostic values. Furthermore, KDM5C and KDM4A were associated with immune cell infiltration in HCC. HDMs were associated with cellular and metabolic processes and may be involved in the regulation of gene expression. Differentially expressed HDM genes identified in NAFLD may provide value to understanding pathogenesis and in the development of epigenetic therapeutic targets. However, on the basis of the inconsistent results of in vitro studies, future in vivo experiments combined with transcriptomic analysis are needed for further validation. Full article

RNA regulates various biological processes, such as gene regulation, RNA splicing, and intracellular signal transduction. RNA’s conformational dynamics play crucial roles in performing its diverse functions. Thus, it is essential to explore the flexibility characteristics of RNA, especially pocket flexibility. Here, we propose a computational approach, RPflex, to analyze pocket flexibility using the coarse-grained network model. We first clustered 3154 pockets into 297 groups by similarity calculation based on the coarse-grained lattice model. Then, we introduced the flexibility score to quantify the flexibility by global pocket features. The results show strong correlations between the flexibility scores and root-mean-square fluctuation (RMSF) values, with Pearson correlation coefficients of 0.60, 0.76, and 0.53 in Testing Sets I–III. Considering both flexibility score and network calculations, the Pearson correlation coefficient was increased to 0.71 in flexible pockets on Testing Set IV. The network calculations reveal that the long-range interaction changes contributed most to flexibility. In addition, the hydrogen bonds in the base–base interactions greatly stabilize the RNA structure, while backbone interactions determine RNA folding. The computational analysis of pocket flexibility could facilitate RNA engineering for biological or medical applications. Full article

Protein–peptide interactions are an essential player in cellular processes and, thus, of great interest as potential therapeutic agents. However, identifying the protein’s interacting surface has been shown to be a challenging task. Here, we present a methodology for protein–peptide interaction identification, implementing phage panning, next-generation sequencing and bioinformatic analysis. One of the uses of this methodology is identification of allergen epitopes, especially suitable for globular inhaled and venom allergens, where their binding capability is determined by the allergen’s conformation, meaning their interaction cannot be properly studied when denatured. A Ph.D. commercial system based on the M13 phage vector was used for the panning process. Utilization of various bioinformatic tools, such as PuLSE, SAROTUP, MEME, Hammock and Pepitope, allowed us to evaluate a large amount of obtained data. Using the described methodology, we identified three peptide clusters representing potential epitopes on the major wasp venom allergen Ves v 5. Full article

Resistance to anoikis is a key characteristic of many cancer cells, promoting cell survival. However, the mechanism of anoikis in hepatocellular carcinoma (HCC) remains unknown. In this study, we applied differentially expressed overlapping anoikis-related genes to classify The Cancer Genome Atlas (TCGA) samples using an unsupervised cluster algorithm. Then, we employed weighted gene coexpression network analysis (WGCNA) to identify highly correlated genes and constructed a prognostic risk model based on univariate Cox proportional hazards regression. This model was validated using external datasets from the International Cancer Genome Consortium (ICGC) and Gene Expression Omnibus (GEO). Finally, we used a CIBERSORT algorithm to investigate the correlation between risk score and immune infiltration. Our results showed that the TCGA cohorts could be divided into two subgroups, with subgroup A having a lower survival probability. Five genes (BAK1, SPP1, BSG, PBK and DAP3) were identified as anoikis-related prognostic genes. Moreover, the prognostic risk model effectively predicted overall survival, which was validated using ICGC and GEO datasets. In addition, there was a strong correlation between infiltrating immune cells and prognostic genes and risk score. In conclusion, we identified anoikis-related subgroups and prognostic genes in HCC, which could be significant for understanding the molecular mechanisms and treatment of HCC. Full article

Microtubules are key players in several stages of the cell cycle and are also involved in the transportation of cellular organelles. Microtubules are polymerized by α/β tubulin dimers with a highly dynamic feature, especially at the plus ends of the microtubules. Therefore, understanding the interactions among tubulins is crucial for characterizing microtubule dynamics. Studying microtubule dynamics can help researchers make advances in the treatment of neurodegenerative diseases and cancer. In this study, we utilize a series of computational approaches to study the electrostatic interactions at the binding interfaces of tubulin monomers. Our study revealed that among all the four types of tubulin-tubulin binding modes, the electrostatic attractive interactions in the α/β tubulin binding are the strongest while the interactions of α/α tubulin binding in the longitudinal direction are the weakest. Our calculations explained that due to the electrostatic interactions, the tubulins always preferred to form α/β tubulin dimers. The interactions between two protofilaments are the weakest. Thus, the protofilaments are easily separated from each other. Furthermore, the important residues involved in the salt bridges at the binding interfaces of the tubulins are identified, which illustrates the details of the interactions in the microtubule. This study elucidates some mechanistic details of microtubule dynamics and also identifies important residues at the binding interfaces as potential drug targets for the inhibition of cancer cells. Full article

Neurodegenerative diseases, tauopathies, constitute a serious global health problem. The etiology of these diseases is unclear and an increase in their incidence has been projected in the next 30 years. Therefore, the study of the molecular mechanisms that might stop these neurodegenerative processes is very relevant. Classification of neurodegenerative diseases using Machine and Deep Learning algorithms has been widely studied for medical imaging such as Magnetic Resonance Imaging. However, post-mortem immunofluorescence imaging studies of the brains of patients have not yet been used for this purpose. These studies may represent a valuable tool for monitoring aberrant chemical changes or pathological post-translational modifications of the Tau polypeptide. We propose a Convolutional Neural Network pipeline for the classification of Tau pathology of Alzheimer’s disease and Progressive Supranuclear Palsy by analyzing post-mortem immunofluorescence images with different Tau biomarkers performed with models generated with the architecture ResNet-IFT using Transfer Learning. These models’ outputs were interpreted with interpretability algorithms such as Guided Grad-CAM and Occlusion Analysis. To determine the best classifier, four different architectures were tested. We demonstrated that our design was able to classify diseases with an accuracy of 98.41% on average whilst providing an interpretation concerning the proper classification involving different structural patterns in the immunoreactivity of the Tau protein in NFTs present in the brains of patients with Progressive Supranuclear Palsy and Alzheimer’s disease. Full article

To investigate the potential of ginsenosides in treating osteoporosis, ginsenoside compound K (GCK) was selected to explore the potential targets and mechanism based on network pharmacology (NP). Based on text mining from public databases, 206 and 6590 targets were obtained for GCK and osteoporosis, respectively, in which 138 targets were identified as co-targets of GCK and osteoporosis using intersection analysis. Five central gene clusters and key genes (STAT3, PIK3R1, VEGFA, JAK2 and MAP2K1) were identified based on Molecular Complex Detection (MCODE) analysis through constructing a protein–protein interaction network using the STRING database. Gene Ontology (GO) analysis implied that phosphatidylinositol-related biological process, molecular modification and function may play an important role for GCK in the treatment of osteoporosis. Function and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis suggested that the c-Fms-mediated osteoclast differentiation pathway was one of the most important mechanisms for GCK in treating osteoporosis. Meanwhile, except for being identified as key targets based on cytoHubba analysis using Cytoscape software, MAPK and PI3K-related proteins were enriched in the downstream of the c-Fms-mediated osteoclast differentiation pathway. Molecular docking further confirmed that GCK could interact with the cavity on the surface of a c-Fms protein with the lowest binding energy (−8.27 Kcal/moL), and their complex was stabilized by hydrogen bonds (Thr578 (1.97 Å), Leu588 (2.02 Å, 2.18 Å), Ala590 (2.16 Å, 2.84 Å) and Cys 666 (1.93 Å)), van der Waals and alkyl hydrophobic interactions. Summarily, GCK could interfere with the occurrence and progress of osteoporosis through the c-Fms-mediated MAPK and PI3K signaling axis regulating osteoclast differentiation. Full article

The variant enrichment analysis (VEA), a recently developed bioinformatic workflow, has been shown to be a valuable tool for whole-exome sequencing data analysis, allowing finding differences between the number of genetic variants in a given pathway compared to a reference dataset. In a previous study, using VEA, we identified different pathway signatures associated with the development of pulmonary toxicities in mesothelioma patients treated with radical hemithoracic radiation therapy. Here, we used VEA to discover novel pathways altered in individuals exposed to asbestos who developed or not asbestos-related diseases (lung cancer or mesothelioma). A population-based autopsy study was designed in which asbestos exposure was evaluated and quantitated by investigating objective signs of exposure. We selected patients with similar exposure to asbestos. Formalin-fixed paraffin-embedded (FFPE) tissues were used as a source of DNA and whole-exome sequencing analysis was performed, running VEA to identify potentially disrupted pathways in individuals who developed thoracic cancers induced by asbestos exposure. By using VEA analysis, we confirmed the involvement of pathways considered as the main culprits for asbestos-induced carcinogenesis: oxidative stress and chromosome instability. Furthermore, we identified protective genetic assets preserving genome stability and susceptibility assets predisposing to a worst outcome. Full article

Lysine SUMOylation plays an essential role in various biological functions. Several approaches integrating various algorithms have been developed for predicting SUMOylation sites based on a limited dataset. Recently, the number of identified SUMOylation sites has significantly increased due to investigation at the proteomics scale. We collected modification data and found the reported approaches had poor performance using our collected data. Therefore, it is essential to explore the characteristics of this modification and construct prediction models with improved performance based on an enlarged dataset. In this study, we constructed and compared 16 classifiers by integrating four different algorithms and four encoding features selected from 11 sequence-based or physicochemical features. We found that the convolution neural network (CNN) model integrated with residue structure, dubbed ResSUMO, performed favorably when compared with the traditional machine learning and CNN models in both cross-validation and independent tests. The area under the receiver operating characteristic (ROC) curve for ResSUMO was around 0.80, superior to that of the reported predictors. We also found that increasing the depth of neural networks in the CNN models did not improve prediction performance due to the degradation problem, but the residual structure could be included to optimize the neural networks and improve performance. This indicates that residual neural networks have the potential to be broadly applied in the prediction of other types of modification sites with great effectiveness and robustness. Furthermore, the online ResSUMO service is freely accessible. Full article

Osteogenesis imperfecta(OI) is a disease caused by substitution in glycine residues with different amino acids in type I collagen (Gly-Xaa-Yaa)n. Collagen model peptides can capture the thermal stability loss of the helix after Gly mutations, most of which are homotrimers. However, a majority of natural collagen exists in heterotrimers. To investigate the effects of chain specific mutations in the natural state of collagen more accurately, here we introduce various lengths of side-chain amino acids into ABC-type heterotrimers. The disruptive effects of the mutations were characterized both experimentally and computationally. We found the stability decrease in the mutants was mainly caused by the disruption of backbone hydrogen bonds. Meanwhile, we found a threshold value of local hydrogen bonding energy that could predict triple helix folding or unfolding. Val caused the unfolding of triple helices, whereas Ser with a similar side-chain length did not. Structural details suggested that the side-chain hydroxyl group in Ser forms hydrogen bonds with the backbone, thereby compensating for the mutants’ decreased stability. Our study contributes to a better understanding of how OI mutations destabilize collagen triple helices and the molecular mechanisms underlying OI. Full article

The PDZ family has drawn attention as possible drug targets because of the domains’ wide ranges of function and highly conserved binding pockets. The PICK1 PDZ domain has been proposed as a possible drug target because the interactions between the PICK1 PDZ domain and the GluA2 subunit of the AMPA receptor have been shown to progress neurodegenerative diseases. BIO124 has been identified as a sub µM inhibitor of the PICK1–GluA2 interaction. Here, we use all-atom molecular dynamics simulations to reveal the atomic-level interaction pattern between the PICK1 PDZ domain and BIO124. Our simulations reveal three unique binding conformations of BIO124 in the PICK1 PDZ binding pocket, referred to here as state 0, state 1, and state 2. Each conformation is defined by a unique hydrogen bonding network and a unique pattern of hydrophobic interactions between BIO124 and the PICK1 PDZ domain. Interestingly, each conformation of BIO124 results in different dynamic changes to the PICK1 PDZ domain. Unlike states 1 and 2, state 0 induces dynamic coupling between BIO124 and the αA helix. Notably, this dynamic coupling with the αA helix is similar to what has been observed in other PDZ–ligand complexes. Our analysis indicates that the interactions formed between BIO124 and I35 may be the key to inducing dynamic coupling with the αA helix. Lastly, we suspect that the conformational shifts observed in our simulations may affect the stability and thus the overall effectiveness of BIO124. We propose that a physically larger inhibitor may be necessary to ensure sufficient interactions that permit stable binding between a drug and the PICK1 PDZ domain. Full article

COVID-19 which was announced as a pandemic on 11 March 2020, is still infecting millions to date as the vaccines that have been developed do not prevent the disease but rather reduce the severity of the symptoms. Until a vaccine is developed that can prevent COVID-19 infection, the testing of individuals will be a continuous process. Medical personnel monitor and treat all health conditions; hence, the time-consuming process to monitor and test all individuals for COVID-19 becomes an impossible task, especially as COVID-19 shares similar symptoms with the common cold and pneumonia. Some off-the-counter tests have been developed and sold, but they are unreliable and add an additional burden because false-positive cases have to visit hospitals and perform specialized diagnostic tests to confirm the diagnosis. Therefore, the need for systems that can automatically detect and diagnose COVID-19 automatically without human intervention is still an urgent priority and will remain so because the same technology can be used for future pandemics and other health conditions. In this paper, we propose a modified machine learning (ML) process that integrates deep learning (DL) algorithms for feature extraction and well-known classifiers that can accurately detect and diagnose COVID-19 from chest CT scans. Publicly available datasets were made available by the China Consortium for Chest CT Image Investigation (CC-CCII). The highest average accuracy obtained was 99.9% using the modified ML process when 2000 features were extracted using GoogleNet and ResNet18 and using the support vector machine (SVM) classifier. The results obtained using the modified ML process were higher when compared to similar methods reported in the extant literature using the same datasets or different datasets of similar size; thus, this study is considered of added value to the current body of knowledge. Further research in this field is required to develop methods that can be applied in hospitals and can better equip mankind to be prepared for any future pandemics. Full article

RNA–protein complexes regulate a variety of biological functions. Thus, it is essential to explore and visualize RNA–protein structural interaction features, especially pocket interactions. In this work, we develop an easy-to-use bioinformatics resource: RPpocket. This database provides RNA–protein complex interactions based on sequence, secondary structure, and pocket topology analysis. We extracted 793 pockets from 74 non-redundant RNA–protein structures. Then, we calculated the binding- and non-binding pocket topological properties and analyzed the binding mechanism of the RNA–protein complex. The results showed that the binding pockets were more extended than the non-binding pockets. We also found that long-range forces were the main interaction for RNA–protein recognition, while short-range forces strengthened and optimized the binding. RPpocket could facilitate RNA–protein engineering for biological or medical applications. Full article