Fig 1.
Analysis of the global DNA-methylation profile in multistage cSCC.
A) Representation of the Principal Component Analysis (PCA) on the methylation data. Centroids of each group of patients are represented by different shapes and colors (legend indicated in B). B) Heatmap with a random sample of 5000 CpGs from the results of the beta regression analysis for differential methylation among all cutaneous squamous cell carcinoma groups of patients. Z-score colour scale ranges from green for lower methylation to red for higher methylation levels.
Fig 2.
Identification of a DNA-methylation signature characterizing the different stages of cSCC.
Heatmap with the results of the elastic net analysis for discrimination among all cutaneous squamous cell carcinoma groups of patients. Rows (CpGs) and columns (individuals) are ordered according to the results of a hierarchical clustering algorithm. Z-score colour scale ranges from green for lower methylation to red for higher methylation levels.
Fig 3.
Comparison of DNA-methylation pattern between low-risk and high-risk stages.
A) Heatmap with the results of the elastic net analysis for differential methylation between low-risk initial invasive and high risk non-metastatic and metastatic cutaneous squamous cell carcinoma (cSCC) stages. Z-score colour scale ranges from green for lower methylation to red for higher methylation levels. B) DNA methylation median β-values of the different epigenomic substructures in the different groups of sCCC. Box plots indicate highly significant (p < 0.001, two-sided t-test) hypermethylation of all substructures in both high-risk non-metastatic and metastatic groups compared to low-risk initial invasive cSCC samples. C) Distribution of selected differentially methylated CpGs in high-risk compared to low-risk cSCC (hypomethylated in green and hypermethylated in red) and total CpG in the Illumina DNA methylation EPIC array in genes, promoters, gene body or intergenic regions. D) Distribution of differentially methylated CpGs in CpG islands, shore, shelf or open sea.
Fig 4.
Analysis of genes associated to differentially methylated CpG between low-risk and high-risk stages.
Hierarchical clustering tree of the enriched gene sets obtained from the GO analysis of the genes selected in the comparison between low-risk initial invasive and high risk non-metastatic and metastatic cutaneous squamous cell carcinoma (cSCC) stages.
Fig 5.
Survival model based on a prognosis DNA-methylation signature.
Representation of the predicted survival curves in patients with cSCC for different estimated risk values by the regression model using 32 CpGs. Patients with higher predicted risk levels show sizeably lower overall survival (OS) in comparison with patients with lower predicted risk levels.