The program receives as input a pair of PDB files (or uploaded files in PDB format) and finds "large enough" geometrically similar substructures in both molecules, which can be rigidly superimposed. It returns a set of hypotheses sorted by the size of the superimposed substructures. The method superimposes the C-alpha atoms of the proteins and is sequence-order independent.
Examples: Bacterial electron transferring proteins
1. (Almost identical structures) 1pmy with 1pza
= 116 bits (274), Expect = 8e-26 Identities = 54/120 (45%), Positives
= 75/120 (62%)
2. (more distant in sequence) 1pmy with 1aaj
= 32.1 bits (71), Expect = 2e-06 Identities = 20/78 (25%), Positives
= 42/78 (53%), Gaps = 3/78 (3%)
Result 1: Match size: 78 RMS: 1.44
Rotation: 1.178 -0.059 -2.615 Translation: 30.230 14.864 17.912
Rotations are given in radians for X-Y-Z axes. Rotating space around the X-axis, then around the Y-axis and finally
the Z-axis would give the required rotation. X-Y-Z translation coordinates
are given in Angstrom units.
Note the similar and the different parts of the two proteins.
insertions/deletions in loops.
3. (same fold, distant structures) 1svy with 2vik
Both are gelsolin (actin depolymerizing) proteins from slime mold and chicken, respectively.
that without the structural information, sequence based alignment can be
different than structural alignment!)
Score = 18.8 bits (37), Expect = 0.025
Identities = 23/67 (34%), Positives = 30/67 (44%), Gaps = 12/67 (17%)
Query: 29 VPVPTLSYGNFYEGDCYVLLSTRLTGSGFSYNIHYWLGLNSSQDEQGAAAIYTTQMD-EY
VP+ T S + GDC+ LL LT I+ + G SS E AA +D E
Sbjct: 20 VPLATSSLNS---GDCF-LLDAGLT-------IYQFNGSLSSPQELNLAAEVARAIDAER 68
Query: 88 LGSVAVQ 94
Sbjct: 69 LGLPLVE 75
To watch only the structurally aligned part, look at 2vik:18-111, 1svy:250-161 (green color).