Hybrid Hypercube Optimization Search Algorithm and Multilayer Perceptron Neural Network for Medical Data Classification

Boosting Whale Optimizer with Quasi-Oppositional Learning and Gaussian Barebone for Feature Selection and COVID-19 Image Segmentation

Whale optimization algorithm (WOA) tends to fall into the local optimum and fails to converge quickly in solving complex problems. To address the shortcomings, an improved WOA (QGBWOA) is proposed in this work. First, quasi-opposition-based learning is introduced to enhance the ability of WOA to search for optimal solutions. Second, a Gaussian barebone mechanism is embedded to promote diversity and expand the scope of the solution space in WOA. To verify the advantages of QGBWOA, comparison experiments between QGBWOA and its comparison peers were carried out on CEC 2014 with dimensions 10, 30, 50, and 100 and on CEC 2020 test with dimension 30. Furthermore, the performance results were tested using Wilcoxon signed-rank (WS), Friedman test, and post hoc statistical tests for statistical analysis. Convergence accuracy and speed are remarkably improved, as shown by experimental results. Finally, feature selection and multi-threshold image segmentation applications are demonstrated to validate the ability of QGBWOA to solve complex real-world problems. QGBWOA proves its superiority over compared algorithms in feature selection and multi-threshold image segmentation by performing several evaluation metrics.

SUPPLEMENTARY INFORMATION

The online version contains supplementary material available at 10.1007/s42235-022-00297-8.

Identification of the Potential Molecular Mechanism of TGFBI Gene in Persistent Atrial Fibrillation

Background

Transforming growth factor beta-induced protein (TGFBI, encoded by TGFBI gene), is an extracellular matrix protein, widely expressed in variety of tissues. It binds to collagens type I, II, and IV and plays important roles in the interactions of cell with cell, collagen, and matrix. It has been reported to be associated with myocardial fibrosis, and the latter is an important pathophysiologyical basis of atrial fibrillation (AF). However, the mechanism of TGFBI in AF remains unclear. We aimed to detect the potential mechanism of TGFBI in AF via bioinformatics analysis.

Methods

The microarray dataset of GSE115574 was examined to detect the genes coexpressed with TGFBI from 14 left atrial tissue samples of AF patients. TGFBI coexpression genes were then screened using the R package. Using online analytical tools, we determined the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis, Gene Ontology (GO) annotation, and protein-protein interaction (PPI) network of TGFBI and its coexpression genes. The modules and hub genes of the PPI-network were then identified. Another dataset, GSE79768 was examined to verify the hub genes. DrugBank was used to detect the potential target drugs.

Results

In GSE115574 dataset, a total of 1818 coexpression genes (769 positive and 1049 negative) were identified, enriched in 120 biological processes (BP), 38 cellular components (CC), 36 molecular functions (MF), and 39 KEGG pathways. A PPI-network with average 12.2-degree nodes was constructed. The genes clustered in the top module constructed from this network mainly play a role in PI3K-Akt signaling pathway, viral myocarditis, inflammatory bowel disease, and platelet activation. CXCL12, C3, FN1, COL1A2, ACTB, VCAM1, and MMP2 were identified and finally verified as the hub genes, mainly enriched in pathways like leukocyte transendothelial migration, PI3K-Akt signaling pathway, viral myocarditis, rheumatoid arthritis, and platelet activation. Pegcetacoplan, ocriplasmin, and carvedilol were the potential target drugs.

Conclusions

We used microdataset to identify the potential functions and mechanisms of the TGFBI and its coexpression genes in AF patients. Our findings suggest that CXCL12, C3, FN1, COL1A2, ACTB, VCAM1, and MMP2 may be the hub genes.

Cross-Tissue Analysis Using Machine Learning to Identify Novel Biomarkers for Knee Osteoarthritis

Background

Knee osteoarthritis (KOA) is a common degenerative joint disease. In this study, we aimed to identify new biomarkers of KOA to improve the accuracy of diagnosis and treatment.

Methods

GSE98918 and GSE51588 were downloaded from the Gene Expression Omnibus database as training sets, with a total of 74 samples. Gene differences were analyzed by Gene Ontology, Kyoto Encyclopedia of Genes and Genomes pathway, and Disease Ontology enrichment analyses for the differentially expressed genes (DEGs), and GSEA enrichment analysis was carried out for the training gene set. Through least absolute shrinkage and selection operator regression analysis, the support vector machine recursive feature elimination algorithm, and gene expression screening, the range of DEGs was further reduced. Immune infiltration analysis was carried out, and the prediction results of the combined biomarker logistic regression model were verified with GSE55457.

Results

In total, 84 DEGs were identified through differential gene expression analysis. The five biomarkers that were screened further showed significant differences in cartilage, subchondral bone, and synovial tissue. The diagnostic accuracy of the model synthesized using five biomarkers through logistic regression was better than that of a single biomarker and significantly better than that of a single clinical trait.

Conclusions

CX3CR1, SLC7A5, ARL4C, TLR7, and MTHFD2 might be used as novel biomarkers to improve the accuracy of KOA disease diagnosis, monitor disease progression, and improve the efficacy of clinical treatment.