Currently, superpose and write_superposed_pdbs have different behaviors - superpose uses the first protein as the reference to rotate others against, and write_superposed_pdbs uses the set of core indices to rotate. Both are problematic - the first protein may be distant from the rest, and the core indices may be empty for divergent datasets.
Solution:
Use the "central" protein (selected using the Geometricus similarity matrix as the protein with a reasonable level of similarity to the rest) as the reference to superpose
Print a warning message/error message when no core indices are found as this could represent a dataset with too divergent proteins for meaningful alignment or with multiple groups of proteins which should ideally be aligned separately.
Currently,
superposeandwrite_superposed_pdbshave different behaviors -superposeuses the first protein as the reference to rotate others against, andwrite_superposed_pdbsuses the set of core indices to rotate. Both are problematic - the first protein may be distant from the rest, and the core indices may be empty for divergent datasets.Solution:
Related to Issue #8