Re: [gmx-users] PCA analysis results

2015-12-18 Thread Tsjerk Wassenaar
Hi Nithyanan, It's guessing a bit, as the actual answer can only be obtained by tracking some parameters during the simulation, but I would suggest to aim for at least 200 ns. It think the relaxation will take more time than you have, so say until (at least) 100 ns, and then you need to have enoug

Re: [gmx-users] PCA analysis results

2015-12-18 Thread Nithyanan Annamalai
Dr Tsjerk, Thank you for your explanation. The PDB model I used in this simulation was obtained from I-TASSER. There is no ligand involved. Dr, may I know what would be the optimal/ suggested simulation time for this protein with 567 aa? On Sat, Dec 19, 2015 at 4:06 AM, Tsjerk Wassenaar wrote:

Re: [gmx-users] PCA analysis results

2015-12-18 Thread Tsjerk Wassenaar
Hi Nithyanan, Right. So now we have that clear, I've had a look at the PDB file :p What you see is the result of filtering over a single eigenvector. You protein is kind of curling up through rotations of domains. You can't describe these rotations with a single eigenvector. They will be dispersed

Re: [gmx-users] PCA analysis results

2015-12-18 Thread Nithyanan Annamalai
Hi Dr Tsjerk, I didn't do any clustering or make it whole. I just use the .xtc that I ontained after the simulation since I found out there is no pbc effect of the protein. On Sat, Dec 19, 2015 at 3:42 AM, Tsjerk Wassenaar wrote: > Hi Nithyanan, > > How did you preprocess the trajectory? Did y

Re: [gmx-users] PCA analysis results

2015-12-18 Thread Tsjerk Wassenaar
Hi Nithyanan, How did you preprocess the trajectory? Did you cluster and/or make everything whole? By the way, 50 ns is probably too short for a protein of that size. It will probably still be relaxing and the eigenvectors will be those of the relaxation. Cheers, Tsjerk On Fri, Dec 18, 2015 at

[gmx-users] PCA analysis results

2015-12-18 Thread Nithyanan Annamalai
Dear GROMACS users, I have performed 50ns MD for 567 aa monomer protein using gromos53a6 ff in explicit water in the cubic box with 2fs time step. I have used g_covar and g_anaeig to perform Principal Component Analysis/ Essential Dynamics analysis. But when I view the animation of the traject