[gmx-users] dPCA "atoms"

[David van der Spoel]

Yang Ye wrote:
> As what we normally do with dPCA, after obtaining the 2D projection of 
> first and second principle, we make it into a 3D plot of free energy 
> surface. The awk program is provided below to do such conversion. On 
> this surface, we can located regions with low free energy and this leads 
> us to go back to 1D projection to locate which frames are inside the 
> regions.  Clustering or averaging may apply to all the frames found in 
> one region. So, in the end, structure are found inside the trajectory 
> for closer inspection.
> 
> BEGIN {
>    i=0
>    grid_num=40.0
> 
>    x_max=0.0
>    y_max=0.0
>    x_min=0.0
>    y_min=0.0
> 
>    temp=300
>    dkt=8.31/1000.0*temp/4.182
> }
> 
> {
>    x[i]=$1
>    y[i]=$2
> 
>    if ($1>x_max) x_max=$1
>    if ($2>y_max) y_max=$2
>    if ($1<x_min) x_min=$1
>    if ($2<y_min) y_min=$2
>    i++
> }
> 
> END {
>    pt_num=i
> 
>    grid_x_size=(x_max-x_min)/grid_num
>    grid_y_size=(y_max-y_min)/grid_num
> 
>    dm=0
>    for(i=0;i<pt_num;i++) {
>       grid[(int((x[i]-x_min)/grid_x_size)), 
> (int((y[i]-y_min)/grid_y_size))]++
>       g=grid[(int((x[i]-x_min)/grid_x_size)), 
> (int((y[i]-y_min)/grid_y_size))]
>       if(dm<g) { dm=g; }
>    }
> 
>    mm=0;
>    for(i=0;i<grid_num;i++) {
>       for(j=0;j<grid_num;j++) {
>          if(grid[i,j]!="") { grid[i,j]=dkt*log(dm/grid[i,j]); if 
> (grid[i,j]>mm) {mm=grid[i,j];} }
>       }
>    }
> 
>    for(i=0;i<grid_num;i++) {
>       for(j=0;j<grid_num;j++) {
>          if(grid[i,j]=="") { grid[i,j]=mm+0.01 }
>          print x_min+(i+0.5)*grid_x_size " " y_min+(j+0.5)*grid_y_size " 
> " grid[i,j]
>       }
>    }
> }
> 
you can also use g_sham to make energy landscapes. however you still 
have to make the connection to 3D space which is not all trivial. 
potential barriers in dihedral space may be trivial in 3D space due to 
vanderwaals interactions. please do also use normal PCA and compare 
results, in particular the interpretation in terms of structure.

-- 
David.
________________________________________________________________________
David van der Spoel, PhD, Assoc. Prof., Molecular Biophysics group,
Dept. of Cell and Molecular Biology, Uppsala University.
Husargatan 3, Box 596,  	75124 Uppsala, Sweden
phone:	46 18 471 4205		fax: 46 18 511 755
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://folding.bmc.uu.se
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++