International Journal of Molecular Medical Science, 2024, Vol.14, No.5, 274-292 http://medscipublisher.com/index.php/ijmms 281 Table 1 Summary of multi-staged integration tools and meta-dimensional integration tools (Adapted from Sathyanarayanan et al., 2019) Tool Omics data Input Output Method References CNAmet CN, ME and GE Gene-level CN and/or ME as binary matrices and GE matrix Weights, scores and associated P-values, and FDR for each gene Independent association of the omics using signal-to-noise ratio weights followed by combining weights to obtain a score Louhimo and Hautaniemi , 2011 iGC CN and GE Gene-level segmented or thresholded CN and GE matrices List of genes for which expression is driven by amplification and deletion with associated P- and FDR values Student’s t-test with unequal variance Lai et al., 2017 PLRS CN and GE Gene-level GE, CN, thresholded CN with CN call probabilities (optional) matrices Spline coefficients of the model fitted for the genes, P-values and FDRvalues Piecewise linear regression splines Leday and van de Wiel, 2013 Oncodriv e-CIS CN and GE Gene-level GE matrix and thresholded CN Scores for each gene that represent the bias towards expression dysregulation due to copy number change Estimation of impact of CN change on GE followed by estimation of standard scores of the impact in tumour and normal samples and finally combining the standard scores Tamborero et al., 2013 MethylM ix ME and GE Probe-level or gene-level ME and gene-level GE matrices List of transcriptionally predictive and differentially methylated genes Linear regression followed by modelling using beta mixture model Gevaert, 2015 SNF Any omics Matrices of the omics data Cluster assignments of samples Weighted similarity network fusion Wang et al., 2014 BCC GE, ME, miRNA, proteomics Matrices of the omics data Adherence values and cluster assignments of samples for multi-omics clustering and individual omics clustering Bayesian clustering based on Dirichlet mixture model Lock and Dunson, 2013 iClusterP lus Anyomics Matrices of the omics data Cluster assignments of samples and multi-omics gene signature Joint modelling followed by feature selection using lasso Mo and Shen, 2018 mixOmic s Anyomics Matrices of the omics data and the labels associated with samples Prediction of class labels for test data and multi-omics gene signature Sparse generalized canonical correlation analysis. Feature selection using lasso Rohart et al., 2017 Note: CN, copy number; GE, gene expression; ME, methylation 5.3 Epigenetic changes 5.3.1 DNA methylation DNA methylation patterns are critical in the regulation of gene expression and have been identified as significant biomarkers in colon cancer. Dysregulated methylation at the single-cell level has been linked to the expression of marker genes that are specific to different colon cancer subtypes (Sun et al., 2021). Huang et al. (2022) mentioned 21 5mC regulatory factors such as TET3 and DNMT1 in their study. And our analysis showed that the 5mC regulatory factors expression was significantly different in normal and tumor samples (Figure 3). High methylation of the O6-methylguanine-DNA methyltransferase (MGMT) gene promoter has been observed in the normal tissues of CRC patients, and it is associated with mutations in p53 and KRAS, indicating its relevance to CRC progression (Guo et al., 2017). Other high-methylation genes reported in normal tissues of CRC patients include SFRP2, TFPI2, NDRG4, BMP3, and ADAMTS14 (Alonso et al., 2015). He et al. (2018) found that APC2 is in a hypermethylated state and could serve as a tumorigenic biomarker for CRC patients in China (He et al.,
RkJQdWJsaXNoZXIy MjQ4ODYzNQ==