Accounting for immunoprecipitation efficiencies in the statistical analysis of ChIP-seq data.

摘要: ImmunoPrecipitation (IP) efficiencies may vary largely between different antibodies and repeated experiments with the same antibody. These differences have a large impact on quality of ChIP-seq data: more efficient experiment will necessarily lead to higher signal background ratio, therefore an apparent larger number enriched regions, compared less experiment. In this paper, we show how IP can be explicitly accounted for in joint statistical modelling data. We fit latent mixture model eight two proteins, from laboratories where are used proteins. use parameters estimate individual experiments, find that these clearly laboratories, amongst technical replicates lab. When account ChIP efficiency, regions bound than ones, at false discovery rate. A priori knowledge binding sites across also included robust detection differentially among propose multiple data sets. The framework present accounts data, allows jointly, rather individually, leading biological conclusions.