Statistical Approaches for Hepatocellular Carcinoma (HCC) Biomarker Discovery

Abstract

Bio informat ics provides an essential tool for the identification of diseases especially human cancer diseases. Also, the availab ility of the co mplete hu man genome has opened the door for the understanding of these diseases as recent technological advances in functional genomics and proteomics have fuelled interest in identifying the biomarkers of complex diseases such as liver cancer which main ly caused to death. Hepatocellular carcinoma (HCC) is one of the co mmon malignant tumours in the world; the liver cirrhosis is the most important leading cause of it. HCC relates to virus infection, carcinogenic compounds, pollution and genetic factors. This work provides a genomic study that focuses on using bioinformat ics approaches to predict the molecular causes of HCC by the investigation of the chromosomal aberrations including gain, or loss of the genomic DNA copy number to provide accurate diagnoses of this disease using Comparative genomic hybridizat ion (CGH) arrays. Diagnosis and understanding of the disease processes will provide a potential t reat ment of the disease at an early stage. The aim is to apply two statistical approaches based on a circular binary segmentation (CBS) algorith m and a Bayesian Hidden Markov Model (HMM ) to a number of hu man chromosomes for analysing array CGH data that accounts for the dependence between neighbouring clones in order to identify genome-wide alternations in copy number fro m the genomic data. Results provide a well identification of the aberration regions in hu man chro mosomes that may lead to robust biomarkers for the early detection of human HCC.