AMBER Archive (2006)

Subject: AMBER: rms in principal component analysis

From: Lydia (aether2006_at_gmail.com)
Date: Tue Jan 17 2006 - 14:54:32 CST


  I want to do principal analysis of 1ns MD simulation (center of mass
translational motion has been removed during the MD simulation at periodic
time intervals) of a protein in order to calculate fluctuation on CA atoms
in the lowest frequency mode. I found that for the same simulation data,
having "rms first *" command in the analysis script or not having "rms
first *" command
will render different results. That is, if I use the pca analysis script as
the following:
------------------------------------------------------------------------------------------------------------------------
/soft/linux/pkg/amber8_070104/exe.serial/ptraj ../parm/monP.prmtop
<<script
 trajin ../output/mw.1.nptL.mdcrd
 reference ../parm/monP.prmcrd
 strip :WAT
 strip :Cl-
 image
 rms first *
 matrix mwcovar @CA name mwcovarmat out mwcovarmat.dat
 analyze matrix mwcovarmat out wt_mwevec.pev name mwevecs vecs 15
 analyze modes fluct out wtrmsfluc_1 stack mwevecs beg 1 end 1
go
script
end
------------------------------------------------------------------------------------------------------------------------

  the output wtrmsfluct_1 is :
------------------------------------------------------------------------------------------------------------------------
Analysis of modes: RMS FLUCTUATIONS
  Atom no. rmsX rmsY rmsZ rms
         1 0.143 0.057 0.160 0.222
         2 0.151 0.070 0.183 0.248
         3 0.151 0.055 0.126 0.204
         4 0.150 0.055 0.091 0.184
         5 0.151 0.053 0.048 0.167
         6 0.167 0.061 0.024 0.179
         7 0.182 0.069 0.020 0.196
         8 0.187 0.014 0.094 0.210
         9 0.092 0.019 0.090 0.130
        10 0.092 0.060 0.089 0.141
        11 0.129 0.063 0.000 0.144
        12 0.735 0.082 0.895 1.161
        13 0.903 0.055 1.235 1.531
        14 1.009 0.612 0.317 1.222
        15 0.396 0.061 0.063 0.405
        16 0.274 0.305 0.195 0.454
        17 0.374 0.276 0.580 0.743
        18 0.261 0.021 0.177 0.316
        ..... ......... ......... ......... ..........
--------------------------------------------------------------------------------------------------------------------------

  But if I delete "rms first *" and rerun the pca analysis, the output
wtrmsfluct_1 will be:
-------------------------------------------------------------------------------------------------------------------------------
Analysis of modes: RMS FLUCTUATIONS
  Atom no. rmsX rmsY rmsZ rms
         1 0.014 0.006 0.012 0.019
         2 0.002 0.006 0.008 0.010
         3 0.005 0.002 0.010 0.011
         4 0.005 0.004 0.004 0.008
         5 0.007 0.001 0.015 0.017
         6 0.021 0.001 0.005 0.021
         7 0.039 0.011 0.003 0.041
         8 0.011 0.042 0.006 0.044
         9 0.068 0.072 0.014 0.101
        10 0.056 0.128 0.113 0.179
        11 0.288 0.139 0.135 0.347
        12 0.932 0.046 0.977 1.351
        13 1.145 0.077 1.355 1.775
        14 1.225 0.719 0.529 1.516
        15 0.701 0.248 0.180 0.765
        16 0.451 0.124 0.021 0.469
        17 0.242 0.433 0.415 0.646
        18 0.172 0.067 0.063 0.194
        ..... ......... ......... ......... ..........
-------------------------------------------------------------------------------------------------------------------------------

   Would someone please tell me that if the rms fitting is actually needed
here in the pca analysis?
Any help or suggestion on this is greatly appreciated.
sincerely
jianping

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu