Dr. Nita Parekh and her students presented the following paper at an International Conference on Frontiers in Nutrition, Medical Genomics, and Drug Discovery, at Vignan University, Guntur held from 31 October – 2 November 2022.
- Identifying DNA Methylation Biomarkers for Triple Negative Breast Cancer Using Biologically Optimized Machine Learning Model – Suba S and Dr. Nita Parekh
Research work as explained by the authors:
DNA methylation is one of the most common epigenetic variations that causes modifications in gene expressions. Increased methylation levels within functional promoters have been associated with an increased risk of breast cancer (BC). Variations in methylation patterns across different subtypes of the disease have been associated with diverse prognosis among patients with same type of cancer. Triple Negative Breast Cancers (TNBC) are known to be highly heterogenous and less responsive to treatments compared to Luminal A and Luminal B subtypes. In this work, we propose a simple approach for identifying CpG biomarkers for clustering patient samples based on a few significant CpG sites that exhibit high variance between TNBC and hormone positive breast cancers. The highly variant CpG sites from the differentially methylated regions are first identified and used as features for a support vector machine (SVM) classifier. We show that CpG sites that have high SVM feature importance scores are able to distinguish TNBC cancer samples from other BC subtypes using singular value decomposition (SVD). This suggests that 10 CpG sites thus identified can be used as diagnostic markers and the associated genes can be used for targeted treatments. CpG sites and the genes specific to metabotropic glutamate receptor binding function and Ovarian steroidogenesis, Cortisol synthesis & secretion, Aldosterone synthesis & secretion, Phospholipase D signalling pathway are seen to be significantly involved due to the differential methylation of these sites. The involvement of the genes such as DNM3, GAL3ST3, ZIC5 and RGS17 in separating the different hormone specific groups are analysed to understand their role in stratifying the groups.
- DNA Methylation based subtype classification of Breast Cancer – Sri Lakshmi Bhavani P and Nita Parekh
Research work as explained by the authors:
Breast cancer is the second leading cause of death among women. To understand and treat a multifactorial disease like cancer, one must understand all its multi-dimensional contributing factors. Subtyping involves categorizing cancer into homogeneous groups with similar properties, which is essential for managing the disease. Gene expression profiling is the most studied technique for cancer subtyping. Gene expression regulation is a multi-level process involving genetic variations like copy number variations, single nucleotide variations, and epigenetic variations like DNA methylation, miRNA expression, etc. Hence, understanding the mechanisms leading to differential gene expression profiles would benefit subtype-specific treatment or therapy. In this work we carried out DNA methylation-associated subtyping for breast cancer. For this work, publicly available breast cancer DNA methylation and gene expression datasets have been downloaded from The Cancer Genome Atlas (TCGA). First, differentially methylated regions (DMRs) associated with genic regions have been identified. Next, using the COX proportional hazard model, the most significant CpG sites influencing the patients’ survival is identified. Pathway analysis of the genes corresponding to these CpG sites were found to be associated to few cancer-related pathways, with some of these genes being known tumor suppressors or oncogenes. This resulted in 7 CpG sites associated to 7 genes. Methylation profiles of some of these genes differ across the PAM50 subtypes and suggest their applicability for DNA methylation-based subtype classification.