Here is the sequence logo predicted by the SAM server.
H = helix
E = sheet
C = loop (or coil)
The taller the letter at each position, the higher the probability of that specific secondary structure for that amino acid.
For example, the Valine at residue 132 is highly predicted to be part of a sheet. However, the Valine at residue 192 is predicted to be anything.
How important is to use predicted structure?
Is this prediction very accurate? I mean that predicted sheet can be longer or shorter by 1-2 residues but it is "for sure" there?
I would use them as suggestions and would trust the higher confidence predictions (the tallest letters), but would definitely NOT fix the structure based solely on these predictions.
What these secondary structure predictions tell us is that, most likely, this structure isn't a bunch of helices!
are key words there then :)