Skip to main content
Fig. 3 | Clinical Epigenetics

Fig. 3

From: Identification of DNA methylation-regulated genes as potential biomarkers for coronary heart disease via machine learning in the Framingham Heart Study

Fig. 3

Dimensionality reduction to screen hub genes in the training set. A Part of the differentially methylated and expressed genes had statistically significant correlations between methylation and expression levels. The diameter of each circle was positively correlated with the absolute values of Spearman correlation coefficients. B Information of the potential DNA methylation-regulated genes. The circles from inside to outside were respectively: PPI networks among the genes, p value histogram and logFC scatter plot of the expression difference, p value histogram and delta beta scatter plot of the methylation difference, names and chromosomal locations of representative genes. C GO clustering enrichment results of the potential DNA methylation-regulated genes. The inner loop showed the Spearman correlation coefficients between methylation and expression status. D Tuning parameter selection of the potential DNA methylation-regulated genes by misclassification error in the LASSO model. E LASSO coefficients profiles of the features. F Intersection between LASSO and SelectFpr test. The dash lines indicated the threshold value of SelectFpr test. Genes chosen by both LASSO and SelectFpr were marked in red. G GO enrichment results of ATG7, BACH2, CDKN1B, DHCR24 and MPO

Back to article page