Hi Happypiggy,
At the moment, we don't have a well-established way to calculate deltapsi between two long read datasets in a rigorous enough manner for our satisfaction, therefore, the Majiq-L visualization is currently only available for use with majiq PSI inputs. This is due to being able to show a fair comparison between the short read and long read data. There is a patch in the works for voila, in the next release, which will allow dPSI and HET inputs, but show a "fake" LR dpsi. This will be for convenience of using majiq dPSI inputs with LR data, rather than a rigorous LR dpsi definition at this time.
For the current usage I would recommend using majiq PSI inputs, possibly with a second $ voila view instance running side by side to show your dPSI version as well.
Let me know if it makes sense,
-San