Skip to main content
Fig. 2 | BMC Bioinformatics

Fig. 2

From: reactIDR: evaluation of the statistical reproducibility of high-throughput structural analyses towards a robust RNA structure prediction

Fig. 2

Comparison between the irreproducible discovery rate (IDR) and reactIDR model. a IDR is a method used for the classification of signals into true (high-coverage and reproducible) and false (low-coverage and irreproducible) signals, based on the Gaussian mixture model, by associating the observation with the cumulative joint probability distribution and pseudo-value x. b In reactIDR, IDR is combined with the hidden Markov model of the three latent classes: loop, stem/bg, and unmapped. The start and end of transcripts should be assigned to the unmapped class. These three classes are defined to handle case/control comparisons with the Gaussian mixture copulas for each condition simultaneously, in which each parameter is optimized to the data and reference structure using the expectation-maximization algorithm

Back to article page