Learning an un-supervised – Clustering algorithm Monte Carlo over Consensus Clustering for Genomic Data for Tumor Identification

Abstract

Clustering involves the grouping of similar objects into a set known as cluster. Objects in one cluster are likely to be different when compared to objects grouped under another cluster. Gene expression is the process by which information from a gene is used in the synthesis of a functional gene product. Subgroup classification is a basic task in high-throughput genomic data analysis, especially for gene expression and methylation data analysis. Mostly, unsupervised clustering methods are applied to predict new subgroups or test the consistency with known annotations. To get a stable classification of subgroups, consensus clustering is always performed. It clusters repeatedly with a randomly sampled subset of data and summarizes the robustness of the clustering. When faced with significant uncertainty in the process of making a forecast or estimation, the Monte Carlo Simulation might prove to be a better solution. Monte Carlo3C is a consensus clustering algorithm that uses a Monte Carlo simulation to eliminate overfitting and can reject the null hypothesis when only one cluster is there.