1g_analyze(1) GROMACS suite, VERSION 4.5 g_analyze(1)
2
6 g_analyze - analyzes data sets
7 VERSION 4.5
9
11 g_analyze -f graph.xvg -ac autocorr.xvg -msd msd.xvg -cc coscont.xvg
12 -dist distr.xvg -av average.xvg -ee errest.xvg -bal ballisitc.xvg -g
13 fitlog.log -[no]h -[no]version -nice int -[no]w -xvg enum -[no]time -b
14 real -e real -n int -[no]d -bw real -errbar enum -[no]integrate
15 -aver_start real -[no]xydy -[no]regression -[no]luzar -temp real -fit‐
16 start real -fitend real -smooth real -filter real -[no]power -[no]subav
17 -[no]oneacf -acflen int -[no]normalize -P enum -fitfn enum -ncskip int
18 -beginfit real -endfit real
19
21 g_analyze reads an ascii file and analyzes data sets. A line in the
22 input file may start with a time (see option -time) and any number of
23 y values may follow. Multiple sets can also be read when they are sep‐
24 arated by & (option -n), in this case only one y value is read from
25 each line. All lines starting with and @ are skipped. All analyses
26 can also be done for the derivative of a set (option -d).
27 All options, except for -av and -power assume that the points are
30 equidistant in time.
31 g_analyze always shows the average and standard deviation of each set.
34 For each set it also shows the relative deviation of the third and
35 fourth cumulant from those of a Gaussian distribution with the same
36 standard deviation.
37 Option -ac produces the autocorrelation function(s).
40 Option -cc plots the resemblance of set i with a cosine of i/2 peri‐
43 ods. The formula is: 2 (int0-T y(t) cos(i pi t) dt)2 / int0-T y(t) y(t)
44 dt
45 This is useful for principal components obtained from covariance analy‐
47 sis, since the principal components of random diffusion are pure
48 cosines.
49 Option -msd produces the mean square displacement(s).
52 Option -dist produces distribution plot(s).
55 Option -av produces the average over the sets. Error bars can be
58 added with the option -errbar. The errorbars can represent the stan‐
59 dard deviation, the error (assuming the points are independent) or the
60 interval containing 90% of the points, by discarding 5% of the points
61 at the top and the bottom.
62 Option -ee produces error estimates using block averaging. A set is
65 divided in a number of blocks and averages are calculated for each
66 block. The error for the total average is calculated from the variance
67 between averages of the m blocks B_i as follows: error2 = Sum (B_i -
68 B)2 / (m*(m-1)). These errors are plotted as a function of the block
69 size. Also an analytical block average curve is plotted, assuming that
70 the autocorrelation is a sum of two exponentials. The analytical curve
71 for the block average is:
72 f(t) = sigma sqrt(2/T ( a (tau1 ((exp(-t/tau1) - 1) tau1/t + 1)) +
74 (1-a) (tau2 ((exp(-t/tau2) - 1) tau2/t + 1)))),
76 where T is the total time. a, tau1 and tau2 are obtained by fitting
77 f2(t) to error2. When the actual block average is very close to the
78 analytical curve, the error is sigma*sqrt(2/T (a tau1 + (1-a) tau2)).
79 The complete derivation is given in B. Hess, J. Chem. Phys.
80 116:209-217, 2002.
81 Option -bal finds and subtracts the ultrafast "ballistic" component
84 from a hydrogen bond autocorrelation function by the fitting of a sum
85 of exponentials, as described in e.g. O. Markovitch, J. Chem. Phys.
86 129:084505, 2008. The fastest term is the one with the most negative
87 coefficient in the exponential, or with -d, the one with most negative
88 time derivative at time 0. -nbalexp sets the number of exponentials
89 to fit.
90 Option -gem fits bimolecular rate constants ka and kb (and optionally
93 kD) to the hydrogen bond autocorrelation function according to the
94 reversible geminate recombination model. Removal of the ballistic com‐
95 ponent first is strongly adviced. The model is presented in O.
96 Markovitch, J. Chem. Phys. 129:084505, 2008.
97 Option -filter prints the RMS high-frequency fluctuation of each set
100 and over all sets with respect to a filtered average. The filter is
101 proportional to cos(pi t/len) where t goes from -len/2 to len/2. len is
102 supplied with the option -filter. This filter reduces oscillations
103 with period len/2 and len by a factor of 0.79 and 0.33 respectively.
104 Option -g fits the data to the function given with option -fitfn.
107 Option -power fits the data to b ta, which is accomplished by fitting
110 to a t + b on log-log scale. All points after the first zero or nega‐
111 tive value are ignored.
112 Option -luzar performs a Luzar & Chandler kinetics analysis on output
114 from g_hbond. The input file can be taken directly from g_hbond -ac,
115 and then the same result should be produced.
116
118 -f graph.xvg Input
119 xvgr/xmgr file
120 -ac autocorr.xvg Output, Opt.
122 xvgr/xmgr file
123 -msd msd.xvg Output, Opt.
125 xvgr/xmgr file
126 -cc coscont.xvg Output, Opt.
128 xvgr/xmgr file
129 -dist distr.xvg Output, Opt.
131 xvgr/xmgr file
132 -av average.xvg Output, Opt.
134 xvgr/xmgr file
135 -ee errest.xvg Output, Opt.
137 xvgr/xmgr file
138 -bal ballisitc.xvg Output, Opt.
140 xvgr/xmgr file
141 -g fitlog.log Output, Opt.
143 Log file
144
147 -[no]hno
148 Print help info and quit
149 -[no]versionno
151 Print version info and quit
152 -nice int 0
154 Set the nicelevel
155 -[no]wno
157 View output xvg, xpm, eps and pdb files
158 -xvg enum xmgrace
160 xvg plot formatting: xmgrace, xmgr or none
161 -[no]timeyes
163 Expect a time in the input
164 -b real -1
166 First time to read from set
167 -e real -1
169 Last time to read from set
170 -n int 1
172 Read sets separated by &
173 -[no]dno
175 Use the derivative
176 -bw real 0.1
178 Binwidth for the distribution
179 -errbar enum none
181 Error bars for -av: none, stddev, error or 90
182 -[no]integrateno
184 Integrate data function(s) numerically using trapezium rule
185 -aver_start real 0
187 Start averaging the integral from here
188 -[no]xydyno
190 Interpret second data set as error in the y values for integrating
191 -[no]regressionno
193 Perform a linear regression analysis on the data. If -xydy is set a
194 second set will be interpreted as the error bar in the Y value. Other‐
195 wise, if multiple data sets are present a multilinear regression will
196 be performed yielding the constant A that minimize chi2 = (y - A0 x0 -
197 A1 x1 - ... - AN xN)2 where now Y is the first data set in the input
198 file and xi the others. Do read the information at the option -time.
199 -[no]luzarno
201 Do a Luzar and Chandler analysis on a correlation function and related
202 as produced by g_hbond. When in addition the -xydy flag is given the
203 second and fourth column will be interpreted as errors in c(t) and
204 n(t).
205 -temp real 298.15
207 Temperature for the Luzar hydrogen bonding kinetics analysis
208 -fitstart real 1
210 Time (ps) from which to start fitting the correlation functions in
211 order to obtain the forward and backward rate constants for HB breaking
212 and formation
213 -fitend real 60
215 Time (ps) where to stop fitting the correlation functions in order to
216 obtain the forward and backward rate constants for HB breaking and for‐
217 mation. Only with -gem
218 -smooth real -1
220 If = 0, the tail of the ACF will be smoothed by fitting it to an expo‐
221 nential function: y = A exp(-x/tau)
222 -filter real 0
224 Print the high-frequency fluctuation after filtering with a cosine
225 filter of length
226 -[no]powerno
228 Fit data to: b ta
229 -[no]subavyes
231 Subtract the average before autocorrelating
232 -[no]oneacfno
234 Calculate one ACF over all sets
235 -acflen int -1
237 Length of the ACF, default is half the number of frames
238 -[no]normalizeyes
240 Normalize ACF
241 -P enum 0
243 Order of Legendre polynomial for ACF (0 indicates none): 0, 1, 2 or
244 3
245 -fitfn enum none
247 Fit function: none, exp, aexp, exp_exp, vac, exp5, exp7 or
248 exp9
249 -ncskip int 0
251 Skip N points in the output file of correlation functions
252 -beginfit real 0
254 Time where to begin the exponential fit of the correlation function
255 -endfit real -1
257 Time where to end the exponential fit of the correlation function, -1
258 is until the end
259
262 gromacs(7)
263 More information about GROMACS is available at <http://www.gro‐
265 macs.org/>.
266 Thu 26 Aug 2010 g_analyze(1)