Statistical Analysis of DNA Methylation Data in a Cervical Cancer Study

DNA methylation occurs when methyl groups attach to cytosine bases on DNA segments. Previous studies have established links between specific methylation patterns and many diseases. In this research, statistical methods are employed to test for significant differences in methylation levels between HIV patients with different stages of cervical cancer. DNA methylation levels are measured at cytosines across the genome with Illumina 450K methylation microarrays. After initial pre-processing to eliminate low-quality data, testing was performed at each cytosine site using t-tests and empirical Bayes tests to identify any statistically significant site level methylation differences between the cervical cancer stages. Two region level statistical methods (Bumphunter and DMRcate) were also applied to identify statistically significant regions of interest in the genome. Significant sites or regions that overlap with genes, CpG Islands, or other genomic annotations can help researchers better understand the molecular impact of DNA methylation and its connection to cervical cancer