I'm trying to use Jpred to predict secondary structure for a protein sequence. When I run J-pred, I get a bunch of hits from PDB. I've also noticed these 'hits' are the same name as the templates i use when using Swiss mod. How do i interpret these hits, do I just ignore them and force jpred to predict structure anyway?
That depends on what your are trying to do. Apparently your query sequence is similar to that of proteins with a known 3D-structure. As it says on Jpred's result page (and help), in this case it might be worth looking at these homologues with experimentally determined structures for information on secondary structure. Most likely it will be more accurate than Jpred's prediction.
How much of the secondary structure you can use from the 3D-structure also depends on which parts align between your sequence and the ones in the PDB. You can show the alignments on Jpred's result page by clicking the button below 'Alignment of PDB hits to your sequence'. However, if you are going with annotations from the 3D-structure, it would probably be better to use SWISS-MODEL directly.
If you don't want to see those hits and just force a Jpred prediction, you can tick 'Skip searching PDB before prediction' in the advanced options before submitting your prediction. The hits are shown merely as a convenience, ignoring them doesn't change anything about Jpred's prediction.