IGLA-3D is a modular method for construction of pairwise 3D alignments (superpositions) of protein structures. It consists of the high-level control module and four special-purpose modules:
the first module (IS) constructs an initial superposition of two protein structures or an initial core, the second module (AC) constructs alignment core given proteins superposition, the third module (FM) finds a motion (shift and rotation) that optimally superimposes the alignment core, the forth module (SM) evaluates the similarity of proteins given a 3D alignment of these proteins.

The control module unites the special-purpose modules using one of the standard schemes for iterative 3D alignment algorithms (the scheme choice depends on the type of the IS module):

Scheme (a)
an initial superposition is constructed (IS) alignment core is constructed based on the current superposition (AC) the superposition parameters (shift vector and rotation matrix) are recalculated using the current core (FM) these two steps are iterated (stop conditions are the number of iterations exceeds certain threshold or the new core coincides with the core that was already seen) similarity of proteins is evaluated (SM) using a set of constructed alignments and for each similarity score the optimal alignment with respect to this score is selected from the set of constructed alignments
Scheme (b)
an initial core is constructed (IS) the superposition parameters (shift vector and rotation matrix) are recalculated using the current core (FM) alignment core is constructed based on the current superposition (AC) these two steps are iterated (stop conditions are the number of iterations exceeds certain threshold or the new core coincides with the core that was already seen similarity of proteins is evaluated (SM) using a set of constructed alignments and for each similarity score the optimal alignment with respect to this score is selected from the set of constructed alignments

The benefit of modularity is the fact that a module can be easily replaced by another module that solves the same task. Hence, if different modules are implemented for one task, one can easily find the best combination for his purposes. For example, testing showed that for certain tasks SM modules that use biological information along with standard geometrical information significantly increase the quality of results in comparison with SM modules that use only geometrical information.

The currents version of IGLA-3D web-service uses the fixed special-purpose modules:

• the IS module constructs initial alignment core for two protein structures using comparison of graphs that represent the secondary structures of proteins. The method is based on the ideas of S. Feranchuck
• the AC module applies the Needleman-Wunsch algorithm to the penalty matrix based on the TM score [1]
• the FM module uses an algorithm similar to the well-known Kabsch algorithm [2]
• the SM module evaluates similarity scores that include normed RMSD (i.e., rmsd / sqrt (core_size)), TM-score, Dali elastic score [3]

However, in future versions of the IGLA-3D web service the users will be able to select the special-purpose modules to be used by the control module from a given list - one module for each task

[1] Y. Zhang, J. Skolnick (2005), TM-align: a protein structure alignment algorithm based on the TM-score // Nucleic Acids Research, 33(7): 2302-2309.

[2] W. Kabsch (1978), A discussion of the solution for the best rotation to relate two sets of vectors // Acta Crystallogr. A, 34: 827-828.

[3] L. Holm, C. Sander (1993), Protein Structure Comparison by Alignment of Distance Matrices // J. Mol. Biol., 233: 123-138.