I run full length of sequences and part of sequences with I-TASSER, and got different predication on these share sequences. How to compare these results? Which one would be more reliable? There is no helix structure with full sequences. Thanks for any suggestion.