I am trying to predict a secondary structure of a 35-aa long peptide. So far, I tried Spider, Jpred and PSIpred servers. All of them are based on the PSI-Blast alignment. For my purpose, I need to predict the secondary structure without PSI-Blast alignment, just based on the aminoacid sequence. Or at least, the possibility to avoid some PDB during the alignment. Do you have suggestion/comment on this?