Skip to main content

Normal Modes Analysis & Flexibility

Let's say you have a few structures and you want to characterize the flexibility of their cavities, but you can't run MD or NMA, at least not locally. You can always upload your PDBs at Bahar's server and get the first, say, 20 normal modes and their frequencies and then use ANA to calculate the cavities flexibilities. In this tutorial, we're going to do just that. As always, all the files are available online.

We will be using the same dynein structures we used on the previous tutorial, just to be sure we're not getting the wrong results.

NMA using Bahar's server

We upload lc8.pdb to the server and run the calculations without changing the defaults, just to be quick.

We get an animation of the first normal model like the one on the next image, and a panel of options.

We go straight to Download files and download from the first 2 links, "Small Eigenvalues" and "Slow eigenvectors".

We name the eigenvalues as pre_frequencies_lc8 and the eigenvectors as pre_modes_lc8. These 2 files require minor formatting changes before using them. In both files, we delete the first column with the row numbers since these are not informative. In addtion, the first 6 frequencies, ---corresponding to the translation and rotation of the center of mass---, must be removed from the pre_frequencies_lc8 file. That is, we edit the files pre_modes_lc8 and pre_frequencies_lc8 in order to keep only the first 20 slowest eigenvectors and frequencies.

in the eigenvalues file ---pre_frequencies_lc8---, we delete the first 6 rows since these correspond to the translational and rotational frequencies so they have no real information. We also delete the last 10 rows since, for some reason, the server gave us 20 eigenvectors and 36 eigenvalues. After these modifications, we have 20 eigenvectors (in the properly formatted modes_lc8 file) and 20 eigenvalues in the frequencies_lc8 file. We do the same for the tctex homodimer and get another two files: modes_tctex and frequencies_tctex.


We now have the vectors and the frequencies of both homodimers. Those are the only new files with respect to the previous tutorial. The structures are the same, so the cavities are the same and ANA's configuration files too.

The contents of our running script ./

ANA2 lc8.pdb -c ecf.cfg -M modes_lc8 -F frequencies_lc8

ANA2 lc8.pdb -c edf.cfg -M modes_lc8 -F frequencies_lc8

ANA2 tctex.pdb -c acb.cfg -M modes_tctex -F frequencies_tctex

ANA2 tctex.pdb -c adb.cfg -M modes_tctex -F frequencies_tctex

Let's run our ./ script:

LC8 cavity C
Rigidity: 4.0808210217

LC8 cavity D
Rigidity: 3.8639151884

TcTex cavity C
Rigidity: 1.8122111870

TcTex cavity D
Rigidity: 1.8893806879

And we see the same difference in flexibility we saw before: TcTex1 has more flexible cavities than LC8. But we also notice the rigidities are much lower this time. This is because we only used 20 vectors and 20 frequencies instead of the whole set of collective coordinates (504 for LC8 and 576 for TcTex1), as we did in the previous tutorial. Yet, we still got the same result with the first 20 normal modes.