Hello Noa,
Internally, this means that the program looked for exons in the splicegraph which should have been at the ends of the junction for some LSV. However, no exons were found.
I would guess it's one of two things: Using a splicegraph.sql file from a different build from another experiment (the ".deltapsi.voila" was not derived from the same build as the "splicegraph.sql" file)
or
Something caused zero lsvs in the build, which is often a problem with the gff3 annotation used. In this case you would see a log during the 'majiq build' step, which said "0 lsvs found" ; could you paste that log to rule this issue out?