Because the model is not created by sequence alignment but by pairwise residue probabilities, BetaWrapPro is able to successfully model proteins with very low sequence similarity to the backbone used for modeling.
What it doesn't do:
BetaWrapPro uses very little information about the sequences of known beta-helices and beta-trefoils. It does not compare a sequence to the known beta-helices and beta-trefoils, or any other known structures, except to build a sequence profile (you do have the option of running an additional search against Pfam HMMs, see below. You should definately do these comparisons for any sequences you are interested in, using for example the NCBI's BLAST service. BetaWrap is not a threading program per se, in that it doesn't compare the sequence to any other possible template structures. As a result it is much faster than threading programs, but it won't notice if your sequence, which might make a mediocre beta-helix, would in fact make a fantastic transmembrane beta-barrel. You should consider using some version of threading or profile program for sequences picked out by BetaWrap (for example 3D-PSSM). But don't be concerned if threading doesn't support BetaWrap's prediction (as long as it doesn't find a highly significant alternative hit) -- the threading programs we tried did not do so well in recognizing similarity between many of the known beta-helices.
It produces sequence-structure alignments that are 76% accurate on the known beta-helix structures, and 86% on the beta-trefoils. We note that it seems to perform less well on the pectin methylesterase SCOP family of beta-helices, and the beta-helix alignments of all structures not in this family are 84% accurate overall. We define an alignment to be accurate if it is within three residues of the exact position in the solved structure.
The accurately aligned regions of the predicted beta-helices average less than 2.0 Å RMSD (root mean square deviation) from the solved structures. The sidechain χ1 angles are correct for 61% of these residues, and χ1+2 are correct for 39%. For the beta-trefoils, the backbone RMSD averages 4.5 Å, with 49% of χ1 angles correct and 25% of χ1+2. It is likely that the higher RMSD for the trefoils compared to the helices is due to the greater structural deviaton between families of trefoils predicted by BetaWrapPro, which in turn makes sidechain prediction more difficult. An angle is counted as correct if it is within 40° of the angle in the solved structure.
SCWRL produces an energy score, which is a measure of how well the sidechains fit onto the given backbone. A high energy score (over 200 or 300) may be an indication that the wrap fits poorly onto the backbone, implying that the sequence may not form the motif. It may also indicate that, although the protein does fold into the specific motif, the sequence-structure alignment is inaccurate.
Reading a wrap:
BetaWrapPro provides several ways to view a wrap. One can download the PDB file provided, which contains the sequence-structure alignment in the 3D structure prediction. Both the beta-helix and beta-trefoil wraps are presented as the entire fasta sequence, and the PSIPRED secondary structure prediction, with the predicted supersecondary structure highlighted. In addition, the beta-helix wraps are shown in a stacking view, which highlights how the predicted rungs align. It should be noted that the "." characters in this view do not indicate gaps in the alignment, but are solely to keep the beta strands aligned correctly.
A key to the colors used for each motif is displayed at the bottom of the results page.
Beta-helix wraps are made up of five rungs, each consisting of three beta strands separated by turn regions. Two of the turns can be highly variable, but the turn between strands two and three is nearly always two residues long. A cross-section of a rung (from the PDB structure 1pcl, Pectate Lyase C) is shown here, with the turns and strands labeled.
You may choose to search for either one or both of the template motifs. If you are only interested in one of them, it will be slightly faster to select that fold rather than both.
BetaWrapPro is occasionally fooled by sequences with simple sequence repeats that have similar structure to the beta-helix motif: the hexapeptide repeat family and the leucine rich repeat family. Pfam has generated profile HMMs for both of these families, and we provide the option of searching for these repeats.
The Pfam and HMMER results attach an E-value to sequence hits. This E-value is an estimate of the expected number of hits with equal or better scores, given the number of query sequences. These E-values are estimated empirically by a calibration process involving random sequences.