Our pipeline uses rMATS to quantify alternative splicing events. rMATS detects five different types of events, as in the figure below. When quantifying events, two independent analyses are performed including and excluding, respectively, the reads on target.
Optionally, a noise filter can be applied to reduce the number of low-abundance significant events included in downstream analysis. The filter is based on inclusion and skipping counts; the sums of inclusion and skipping counts for each group of samples must be above a threshold for the event to be retained.
