Preliminary Analysis of Within-Sample Co-methylation Patterns in Normal and Cancerous Breast Samples. (October 2019)

• Record Type: Journal Article
• Authors: Sun, Lillian; Namboodiri, Surya; Chen, Emily; Sun, Shuying
• Is Part Of: Cancer informatics. Volume 18(2019)
• Publication Date: 2019-10
• Subjects: Within-sample co-methylation -- bioinformatics -- breast cancer; Bioinformatics -- Periodicals; Biology -- Data processing -- Periodicals; Cancer -- Periodicals; Cancer -- Research -- Periodicals; Computational biology -- Periodicals; 570.285
• DOI: 10.1177/1176935119880516
• Languages: English
• ISSNs: 1176-9351

Abstract

DNA methylation plays a significant role in regulating the expression of certain genes in both cancerous and normal breast tissues. It is therefore important to study within-sample co-methylation, ie, methylation patterns between consecutive sites in a chromosome. In this article, we develop 2 new methods to compare co-methylation patterns between normal and cancerous breast samples. In particular, we investigate the co-methylation patterns of 4 different methylation states/levels separately. Using these 2 methods, we focus on addressing the following questions: How often does 1 methylation state change to other methylation states and how is this change dependent on chromosome distance? What co-methylation patterns do normal and cancerous breast samples have? Do genomic sites with different methylation states/levels have different co-methylation patterns? Our results show that cancerous and normal co-methylation patterns are significantly different. We find that this difference exists even when the physical distance of 2 sites are less than 50 bases. Breast cancer cell lines tend to remain in the same methylation state more often than normal samples, especially for the no/low or high/full methylation states. We also find that the co-methylation region lengths for various methylation states (no/low, partial, and high/full methylation states) are very different. For example, the co-methylation region lengths for partial methylation regions are shorter than the unmethylated or fully methylated regions. Our research may provide a deep understanding of co-methylation patterns. These co-methylation patterns will aid in discovering and understanding new methylation events that may be related to novel biomarkers.