I use it to measure flexibility amongst families of proteins or within different isoforms or mutants of the same protein. I am guessing that two completely different proteins will have flexibility at different regions. Unless the secondary structure is similar then it would be pointless to do so, but even in using RMSF you have to decide the regions or times in the trajectory when you are measuring. In other words looking at periods right after or during equilibrium would give results indicating heightened flexibility so you would want to probably do the RMSF based on the stable time regions indiacted by the RMSD.

I would say that the answer is yes, but you have to be fairly careful with this. First of all, what do you mean by RMSF? The overall RMSF of the backbone tracked over time? Or the per-residue RMSF averaged over time? And what do you want to say about the flexibility? For instance, if you have two very different proteins, one that is quite compact with lots of secondary structure and short loops, and one that is floppy with large loops and tails, then yes, I think you could use RMSF to broadly say that one protein is more "flexible" than another. But like most things, the devil is in the details. In the case I just described, analyzing RMSF only over regions of regular secondary structure might give you quite a different picture. This is where looking at the RMSF per residue might be more informative. But then you start getting into problems if the proteins are very different, since the different residues mean different things.

I performed a comparison between a mesophilic and a thermophilic protein that were homologous to each other, but with only low sequence identity (maybe 20% or so). To do this, I performed a sequence alignment and identified insertions in the mesophilic protein. Then I displayed the data as RMSF vs aligned residue number and left gaps in the data for protein 1 when protein 2 had an insertion. You can see this work at http://peds.oxfordjournals.org/content/23/5/327.full.pdf (figure 5).

RMSF calculates deviation between the position of Residues. So you can compare the flexibilities of parts of the proteins

Comparison between two proteins can be made by taking other measures in to account like radius of gyration, RMSD.

Thanks for everybody' help. Now I am more clear about RMSF with protein flexibility. Ya, compare need concentrate on details, Some regions of the protein maybe flexible whereas other regions more stiff. So , Ya, from RMSF to compare completely different proteins globle flexiblity are some not so accurate, I will concentrate on compare regions on the same protein.

Great! Thanks for everybody's help and patience. Best wishes!