RESUMO
Background/aim: The complicated nature of tumor formation makes it difficult to identify discriminatory genes. Recently, transcriptome-based supervised classification methods using support vector machines (SVMs) have become popular in this field. However, the inclusion of less significant variables in the construction of classification models can lead to misclassification. To improve model performance, feature selection methods such as enrichment analysis can be used to extract useful variable sets. The detection of genes that can discriminate between normal and tumor samples in the association of cancer and disease remains an area of limited information. We therefore aimed to discover novel and practical sets of discriminatory biomarkers by utilizing the association of cancer and disease. Materials and methods: In this study, we employed an SVM classification method for differentially expressed genes enriched by Disease Ontology and filtered nondiscriminatory features using Wilk's lambda criterion prior to classification. Our approach uses the discovery of disease-associated genes as a viable strategy to identify gene sets that discriminate between tumor and normal states. We analyzed the performance of our algorithm using comprehensive RNA-Seq data for adenocarcinoma of the colon, squamous cell carcinoma of the lung, and adenocarcinoma of the lung. The classification performance of the obtained gene sets was analyzed by comparison with different expression datasets and previous studies using the same datasets. Results: It was found that our algorithm extracts stable small gene sets that provide high accuracy in predicting cancer status. In addition, the gene sets generated by our method perform well in survival analyses, indicating their potential for prognosis. Conclusion: By combining gene sets for both diagnosis and prognosis, our method can improve clinical applications in cancer research. Our algorithm is available as an R package with a graphical user interface in Bioconductor (https://doi.org/10.18129/B9.bioc.SVMDO) and GitHub (https://github.com/robogeno/SVMDO).
RESUMO
Artificial intelligence, machine learning, health care robots, and algorithms for clinical decision-making are currently being sought after in diverse fields of clinical medicine and bioengineering. The field of personalized medicine stands to benefit from new technologies so as to harness the omics big data, for example, to individualize and accelerate cancer diagnostics and therapeutics in particular. In this overarching context, breast cancer is one of the most common malignancies worldwide with multiple underlying molecular etiologies and each subtype displaying diverse clinical outcomes. Disease stratification for breast cancer is, therefore, vital to its effective and individualized clinical care. The support vector machine (SVM) is a rising machine learning approach that offers robust classification of high-dimensional big data into small numbers of data points (support vectors), achieving differentiation of subgroups in a short amount of time. Considering the rapid timelines required for both diagnosis and treatment of most aggressive cancers, this new machine learning technique has important clinical and public applications and implications for high-throughput data analysis and contextualization. This expert review describes and examines, first, the SVM models employed to forecast breast cancer subtypes using diverse systems science data, including transcriptomics, epigenetics, proteomics, and radiomics, as well as biological pathway, clinical, pathological, and biochemical data. Then, we compare the performance of the present SVM and other diagnostic and therapeutic prediction models across the data types. We conclude by emphasizing that data integration is a critical bottleneck in systems science, cancer research and development, and health care innovation and that SVM and machine learning approaches offer new solutions and ways forward in biomedical, bioengineering, and clinical applications.
