Four different models have been generated for each peptide, which include (i) model I with all peptide residues in extended conformation (φ= ψ=180o), (ii) second model II with backbone torsion angles corresponding to regular secondary structure states predicted by BetaTurns, (iii) third model III with torsion angles corresponding to regular secondary states and β-turns and (iv) fourth model IV with side chain χ angles assigned using standard Dunbrack backbone dependent rotamer library in addition to main-chain φ, ψ angles of model III.
It has been found that the model containing β-turns has the least root mean square deviation in comparison to model containing regular secondary structure information alone. The following Table shows the prediction performance in terms of averaged backbone rmsd for all the 4 models.
Table: Averaged backbone root mean deviation before and after energy minimization and dynamics simulations
