I was reading a recent article that mentions that there were two isoform for a particular gene they were interests in. However when I went to ensembl it turns out that there were a lot more including non protein versions. The question is I'm assuming in the ensembl transripts they are all predicted correct? How reliable are they and can I somehow find out if those particular versions has been experimentally confirm? Here is a a random example for the NTRK3 gene which shows here at least 10 transcripts related to this gene. I'm wondering if CCDS can help here or uniprot which has experimental evidence.