Is ensemble NMA analysis of DHFR sequence sensitive?

[Former user]

I was going through this vignette on Bio3d website (http://thegrantlab.org/bio3d/tutorials/ensemble-nma-part-2 ) and was wondering if nma.pdb() is sensitive to the sequence differences between different ca alpha PDBs of DHFR analyzed? If not, then why would modes be different unless structures are different? Are structures different? That would explain differences in the modes across species.

Is there as nma force field in bio3d that captures sequence differences? Would CA model capture such sequence sensitivity?

[Barry Grant]

These are great questions.

The default calpha force-file, originally developed by Konrad Hinsen, employs a spring force constant differentiating between nearest-neighbour pairs along the backbone and all other pairs. It does not consider amino-acid type, so therefore it is insensitive to sequence differences.

You are thinking along the correct lines, namely structure differences leading to mode differences. Can you think of a way to answer your own question "are structures different?". E.g. using the Bio3D functions rmsd(), pca() etc. etc. and then maybe applying clustering with hclust() . The whole point of Bio3D (and these vignettes in particular) is to facilitate folks exploring these types of questions so I am happy to see your post.

Lars has been working on refining these functions recently and we have a new version 2.2 of the package about to appear shortly. I will let him comment on the other force field choices and new things that might be coming.

[Lars Skjærven]

Jepps.. the only dependency of amino acids types in the calpha force field is due to the mass weighting of Hessian matrix. The sdENM force field on the other hand contains specific spring force constants for each pair of amino acid types. This force field should therefore be more sequence dependent, and might be better for these kind of comparative analysis (but it's still a CA model). See help(load.enmff) for more info and the related paper: Dehouck Y. & Mikhailov A.S. (2013) PLoS Comput Biol. To which degree this force field would capture sequence sensitivity I think is an open question. Certainly worth looking into! There is also another force field which places an extra bead at the CB or CG atom of each residue. I think you'll find a reference to it here: http://pubs.acs.org/doi/abs/10.1021/ct400399x