Bayesian posterior probability classification of colorectal cancer probed with Affymetrix microarray technology

Colorectal cancer is one of the most common types of cancer worldwide. Assuming increased or decreased gene expression is the reason for abnormal cells work and processes interference in the colorectal region, in our previous work we used data from Illumina microarray technology to analyse gene expression values. Once we have unveiled biomarker genes and developed methodology for Bayesian posterior probability classification, we proceeded with implementing the same methodology on data obtained from Affymetrix microarray technology. However, our research results showed that different microarray technologies require different statistical approach for classification analyses. In this paper we use colorectal data probed with Affymetrix microarray technology, and propose a new methodology that intends to eliminate the noise and produce more robust preprocessed data appropriate for prior distribution modelling. This allows us to construct an efficient Bayesian a posteriori classificator. In order to test the procedure reliability we used different set of carcinogenic and healthy patients.