Hi,
Yes, you can use the 'filter' command and/or the 'outtraj' command
with the 'maxmin' keyword to extract frames based on specific dataset
values. Of course, you need to ensure that the trajectory you are
extracting frames from is the same one used to generate the covariance
matrix you used. So something like this:
# Step 1: Read in previously calculated PCs.
# I'm assuming PC1 in column 2 and PC2 in column 3
readdata pca12.dat name MyPcProjection
# Step 1: Fit trajectory frames.
parm complex.psf
trajin complex.dcd
trajin complex.dcd
trajincomplex.dcd
rms first .CA
average crdset avg_set
createcrd traj_set
run
crdaction traj_set rms ref avg_set .CA
# Step 2: Use maxmin to separate frames by bin values in MyPcProjection.
# MyPcProjection:2 should be PC1 from column 2.
crdaction traj_set outtraj Bin1.dcd maxmin MyPcProjection:2 min
<min1-1> max <max1-1> maxmin MyPcProjection:3 min <min1-2> max
<max1-2>
...
Here <minX-Y> and <maxX-Y> are the minimum and maximum values for
defining bin X, principal component projection Y. You can do an
'outtraj' command for each bin of interest. Now keep in mind the
frames extracted in such a manner are "representative" only insofar as
they come from that bin.
To make things more efficient in the future (and assuming disk space
is not an issue), it may be worth it to save the fit trajectory (i.e.
the trajectory you use to calculate the covariance matrix and do
projections) to a separate file so you don't have to keep regenerating
it on the fly.
Hope this helps,
-Dan
On Fri, Feb 3, 2023 at 4:46 AM Hocine EL KHAOUDI ENYOURY via AMBER
<amber.ambermd.org> wrote:
> parm complex.psf
> trajin complex.dcd
> trajin complex.dcd
> trajincomplex.dcd
>
> rms first .CA
> average crdset avg_set
>
> createcrd traj_set
>
> run
>
> crdaction traj_set rms ref avg_set .CA
>
> crdaction traj_set matrix covar :.CA name pro_covar out covmat.dat
>
> runanalysis diagmatrix pro_covar out evecs.dat vecs 100 name myevecs nmwiz nmwizvecs 100 nmwizfile complex.nmd nmwizmask :.CA
> runanalysis modes eigenval name myevecs out evalfrac.dat
> runanalysis modes fluct out rmsfluct1-10.dat name
> myevecs beg 1 end 10 #to write the fluctuations of first 10 PCs
>
> runanalysis modes fluct out rmsfluct1-2.dat name myevecs beg 1 end 2
>
> runanalysis modes fluct out rmsfluct-mode1.dat name myevecs beg 1 #to write the fluctuations of first PC
>
> crdaction traj_set projection p1 modes myevecs beg 1 end 20 .CA out pca.dat #to write the projections
> of first 20 PCs
>
> crdaction traj_set projection p2 modes myevecs beg 1 end 2 :.CA out pca12.dat #to write the projections of first 2
> PCs
>
> runanalysis hist p1:1 p1:2 bins 100 free 300 out fhist-all.CA.agr
>
> Is there a way to extract a representative structure of each bin of the plot?
>
>
> Thanks!
>
> Hocine
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Feb 06 2023 - 06:00:03 PST
