Fig. 2
From: omniCLIP: probabilistic identification of protein-RNA interactions from CLIP-seq data
Illustration of omniCLIP application. a Shown is the read coverage of HES1 for two PAR-CLIP and two RNA-seq libraries as well as reads with diagnostic events. Here, the T-C conversions are shown in (red). b Application of omniCLIP. First, the probability of each position and each state is computed using the coverage profile model and the diagnostic event model. Next, the transition probabilities are computed based on the coverage at each position. Finally, a Non-homogeneous Hidden Markov Model is applied to segment the sequence in to peak regions (P) and non-peak regions (N, B1, B2)