|Thesis abstract: |
Chromatin Immunoprecipitation (ChIP) followed by massively parallel DNA sequencing, commonly known as ChIP-seq, reveals splay biological aspects including chromatin modifications and DNA-protein in vivo interactions by identifying the binding sites of DNA-associated proteins. However, due to experimental obstacles, the signal of biding patterns are not purely unbiased. Comparative ChIP-seq data analysis is a possible solution. We utilize the single test statistic inferred from data fusion of extreme value probabilities (i.e. p-values) of intersection binding sites across Biological/Technical replicates (where similar biding pattern is expected), as an appropriate mean of assessing the veracity of binding sites. The challenges include determination of an adequate data fusion method, an asymptotically optimal algorithm for intersecting binding¿s determination and finally, the verification and interpretation of obtained results.