“align” performs a sequence alignment followed by a structural
superposition, and then carries out zero or more cycles of
refinement in order to reject outliers.
align mobile, target [, object=name ]
mobile = string: atom selection for mobile atoms
target = string: atom selection for target atoms
object = string: name of alignment object to create
align protA////CA, protB////CA, object=alnAB
If object is specified, then align will create an object which
indicates paired atoms and supports visualization of the alignment
in the sequence viewer.
Q & A
Please describe/reference the alignments algorithms used in the Pymol alignment. Does the alignment include side chains?
In short, “align” is a automated multi-step superposition algorithm based on dynamic programming and iterative refinement. “align” first performs a per-residue global dynamic-programming sequence alignment for the input atom selections using the BLOSUM62 weightings from BLAST. Then, a per-atom correspondance is established between atoms in the selections. Matching side chains atoms will be included if they were provided in the selection arguments. An initial superposition is then performed followed by up to five cycles of iterative refinement wherein atoms with per-atom deviations over two standard deviations from the mean deviation (if any) are thrown out and the fit is repeated. Finally, the number of atoms remaining at the end of this procedure is printed out along with the resulting RMS value for those atoms.
CAUTION: Because the output RMS value and atom count are the result of a heuristic algorithm, these values should not be used for comparing relative similarity across multiple pairs of structures.