I constructed a PDB file of main-chain from its CA-trace using pd2_ca2main
app. I used TMalign
to compute RMSD values from those two files.
The research paper of TMaligh says that:
"TM-align only employs the backbone Cα coordinates of the given protein structures; however, the methodology is readily generalized to any type of atom."
Therefore, the source PDB and the resultant PDB have the same CAs. Right? If so, why do I see a deviation when I compare the source PDB and the resultant PDB?