Module-based knowledge discovery for multiple-cytosine-variant methylation profile
Soft Computing for Biological Systems
Methylation-based study is currently a popular ongoing research topic. The researchers generally use 5-methylcytosine (5-mC) samples for their study since this category of samples is the highest stable methylation cytosine variant, and the impact of 5-mC methylation on different diseases is known to the common people. But, through recent studies, it has been observed that other cytosine variants (e.g., 5-hmC) have also high impact on those diseases. Therefore, in this chapter, we firstly demonstrate the abovementioned different cytosine variants. In the second part of the chapter, we describe a framework of identifying co-methylated gene modules on a methylation profile having multiple cytosine variants (viz., 5-hmC and 5-mC samples). For this, at first we determine significant genes using statistical method. Thereafter, weighted topological overlap matrix (weighted TOM) measure and average linkage method are applied, consecutively on the resultant significant genes. Then dynamic tree cut method with color thresholding is utilized, and co-methylated gene modules are identified from it. The resultant gene modules are then validated biologically by KEGG pathway and gene ontology analyses. Moreover, regulatory transcription factors (TFs) and targeter miRNAs connected with the genes belonging to the different modules are found, and further biological validation has been carried out on them. Finally, other related module-based and correlation-based popular computation methodologies and applications are also shortly demonstrated.
Mallik, Saurav and Maulik, Ujjwal, "Module-based knowledge discovery for multiple-cytosine-variant methylation profile" (2018). Book Chapters. 12.