• Accessibility and secondary structure assignment
• Score of an individual position
• Score of a match
• Benchmark
• Score calibration

Accessibility and secondary structure assignment

The solvent accessibility and secondary structure values are collected from DSSP files. For the solvent exposure of a residue, a relative (normalized) value is calculated as the ratio of the residue's accessibility DSSP value to the residue accessible surface area value as defined by Miller and co-workers (Miller et al., 1987). The latter is calculated for the residue in a Gly-Xaa-Gly tripeptide in extended conformation. The relative accessibility varies between 0 and 1. The secondary structure assignments are: H = alpha helix, B = residue in isolated beta-bridge, E = extended strand (participates in beta ladder), G = 3-helix (3/10 helix), I = 5 helix (pi helix), T = hydrogen bonded turn and S = bend.

Score of an individual position

Based on the structural study of true motifs, accessibility (S_A (p)) and secondary structure (S_SSE (p)) score of a position p are assigned as follows:

Accessibility:

             relative accessibility >= 0.7: S_A (p) = 1
             relative accessibility < 0.7: S_A (p) = relative accessibility

             helix: S_SSEM(p) = 0.3
             strand: S_SSE (p) = 0.5
             G-helix: S_SSE (p) = 0.7
             loop: S_SSE (p) = 1.0

Score of a match

Given a motif match that can be modeled onto a structure domain and such that len(match) = N, its global S_A and S_SSE scores are evaluated as:

             S_A (match) = ∑p (S_A (p))/N
             S_SSE (match) = ∑p (S_SSE (p))/N

Benchmark

The benchmark consists of the whole set of ELM's true motifs that can be mapped onto a domain structure (at ≥ 70% sequence similarity). A set of reliable false positives (FPs) is determined as well. Our benchmark is composed of 218 TPs and 28790 FPs.

Score calibration

The percentage of TPs and FPs has been calculated for 0.1 S_A and S_SSE score bins and plotted (figure 1 and 2).

Figure 1 - TPs and FPs frequencies versus S_A score bins

Figure 2 - TPs and FPs frequencies versus S_SSE score bins

Based on these findings, we assigned the score range into three categories:

range	colour	accessibility and 2D structure conditions
poor context	grey	(0.0 ≤ SA< 0.3) or (0.3 ≤ SA ≤ 0.4 and 0.0 ≤ SSSE < 0.5)
quite good context	half blue, half grey	(0.3 ≤ SA ≤ 0.4 and 0.5 ≤ SSSE ≤ 1.0) or (0.4 < SA ≤ 0.7 and 0.0 ≤ SSSE ≤ 0.5)
best context	blue	(0.4 < SA ≤ 0.7 and 0.5 < SSSE ≤ 1.0) or (SA ≥ 0.7)

By evaluating the percentage of TPs and FPs into these categories, we obtained the results in figure 3.


Figure 3 - TPs and FPs frequencies versus range categories.