RMSD: fit = T or fit = F?

[NANCY D'ARMINIO]

Hello, I’m using the RMSD function to calculate the RMSD between two ligands. I’m analizing two pdb files containing the protein and the ligand. I’m also comparing results obtained from R with the ones from Discovery and I got different values. More specifically, for the first ligand, I I used fit = F and the result was near to the one from Discovery (using fit = T results are completely different). For the second ligand, instead, the situation is on the opposite side, fit = T is similar to Discovery and fit = F is far. I’ve read that “fit” applies Kabsch algorithm, but how I can understand when applying it or not? I have to calculate RMSD on large scale and I can’t do it on Discovery manually. Thank you!

‌

[Xinqiu Yao]

Hi,

In principle, you should alway fit structures before calculating RMSD. How to fit exactly (e.g., based on C-alpha or all-heavy atoms, what reference to choose, etc.) is of course dependent on the specific question. But remember, rmsd() returns a result for whatever you used for fitting. So, for example, if you want to fit the structures based protein C-alpha atoms and want to calculate RMSD of ligands, then a better solution would be calling fit.xyz() for fitting first and use the result to call rmsd() with fit=FALSE . For example,

b.fitted <- fit.xyz(a, b, ca.inds.a, ca.inds.b)
rmsd(a, b.fitted, lig.inds.a, lig.inds.b, fit=FALSE)

‌