Abstract:
|
Hi-C chromosome conformation data provides valuable insights on the organization of chromatin in the cell, which may reveal mechanisms for regulation and expression. Analysing Hi-C data is difficult due to high variability and numerous biases. We introduce a new Hi-C probabilistic model to expand upon existing methods for detecting chromatin interactions. To improve detection of small scale functional interactions, we incorporate larger chromatin structures into our hierarchical model, to prevent these structures from distorting estimates of baseline contact rates. In addition, when there are replicate samples this method accounts for technical variation in the Hi-C procedure producing improved peak estimation.
|