Figure 2From: Reduced representation of protein structure: implications on efficiency and scope of detection of structural similarityUsing the geometric criterion (Eq. 4) exclusively to detect the query in PDB25. The histogram shows the cumulative percentage of cases for which the self-match score is found within the top percentage shown on the x-axis. Full line: using δ = 0.1, and dashed line: using δ = 0.3. For both choices of δ self match is always the highest on the list if the sequential ordering (Eq. 6) is used as an additional requirement (i.e. the graph looks the same as δ = 0.1 case shown in red here). Thus in a general comparison strategy one cannot rely on matching directions alone - additional constraints are needed to reduce the number of false positives.Back to article page