Statistical Methods for Genome-wide Tiling Microarray and its Application to ChIP-chip Experiments
Array based Chromatin-immunoprecipitiation (ChIP-chip) method is a widely used positional analysis to obtain direct and physical binding information where the transcription factors, nucleosomes and many other key proteins interact with the genome. The binding sites obtained from ChIP-chip analysis provide a mechanistic understanding of the complex programming of gene regulation. However, the genome-wide tiling array, with 35bp spacing and about 45 millions probes to cover entire human genome of 3 billion base-pairs, poses a computational and statistical challenge for the accurate and efficient data analysis. In the presentation, we will discuss the methods currently available to the community for data normalization, binding site detection, cis-regulatory element identification for a transcription factor binding, and statistical properties of binding motifs. Data integration with gene expression, siRNA knockdown will also be presented. An unique binding site/gene set enrichment analysis method with application to breast cancer study will be discussed in the presentation.