[gmx-users] Regarding root mean square inner product from PCA

[vijayakumar gosu]

Dear Gromacs users,

I have done simulations of wild type and mutant proteins for 200ns. Further
for the backbone of the whole MD trajectory i have performed PCA analysis,
then calculated RMSIP for first 10 eigen vectors by using the wild-type and
mutant eigenvectors. When i look at the overlap of the variance, between WT
and mutant, third eigen vectors are identical. however when i checked for
the correlated motions along third eigen vector, using porcupine plot,  i
could see the difference in the direction of motion as well as in the
magnitude between WT and mutant.  (i have used -extr flag  to extract the
extreme conformation with 30 frames). I am bit confused to interpret the
result. any suggestions please.

Thanks in advance

Vijayakumar Gosu