1GMX-ANALYZE(1) GROMACS GMX-ANALYZE(1)
2
6 gmx-analyze - Analyze data sets
7
9 gmx analyze [-f [<.xvg>]] [-ac [<.xvg>]] [-msd [<.xvg>]] [-cc [<.xvg>]]
10 [-dist [<.xvg>]] [-av [<.xvg>]] [-ee [<.xvg>]]
11 [-fitted [<.xvg>]] [-g [<.log>]] [-[no]w] [-xvg <enum>]
12 [-[no]time] [-b <real>] [-e <real>] [-n <int>] [-[no]d]
13 [-bw <real>] [-errbar <enum>] [-[no]integrate]
14 [-aver_start <real>] [-[no]xydy] [-[no]regression]
15 [-[no]luzar] [-temp <real>] [-fitstart <real>]
16 [-fitend <real>] [-filter <real>] [-[no]power]
17 [-[no]subav] [-[no]oneacf] [-acflen <int>]
18 [-[no]normalize] [-P <enum>] [-fitfn <enum>]
19 [-beginfit <real>] [-endfit <real>]
20
22 gmx analyze reads an ASCII file and analyzes data sets. A line in the
23 input file may start with a time (see option -time) and any number of
24 y-values may follow. Multiple sets can also be read when they are sep‐
25 arated by & (option -n); in this case only one y-value is read from
26 each line. All lines starting with # and @ are skipped. All analyses
27 can also be done for the derivative of a set (option -d).
28 All options, except for -av and -power, assume that the points are
30 equidistant in time.
31 gmx analyze always shows the average and standard deviation of each
33 set, as well as the relative deviation of the third and fourth cumulant
34 from those of a Gaussian distribution with the same standard deviation.
35 Option -ac produces the autocorrelation function(s). Be sure that the
37 time interval between data points is much shorter than the time scale
38 of the autocorrelation.
39 Option -cc plots the resemblance of set i with a cosine of i/2 periods.
41 The formula is:
42 2 (integral from 0 to T of y(t) cos(i pi t) dt)^2 / integral from 0 to T of y^2(t) dt
44 This is useful for principal components obtained from covariance analy‐
46 sis, since the principal components of random diffusion are pure
47 cosines.
48 Option -msd produces the mean square displacement(s).
50 Option -dist produces distribution plot(s).
52 Option -av produces the average over the sets. Error bars can be added
54 with the option -errbar. The errorbars can represent the standard
55 deviation, the error (assuming the points are independent) or the
56 interval containing 90% of the points, by discarding 5% of the points
57 at the top and the bottom.
58 Option -ee produces error estimates using block averaging. A set is
60 divided in a number of blocks and averages are calculated for each
61 block. The error for the total average is calculated from the variance
62 between averages of the m blocks B_i as follows: error^2 = sum (B_i -
63 <B>)^2 / (m*(m-1)). These errors are plotted as a function of the
64 block size. Also an analytical block average curve is plotted, assum‐
65 ing that the autocorrelation is a sum of two exponentials. The analyt‐
66 ical curve for the block average is:
67 f(t) = sigma``*``sqrt(2/T ( alpha (tau_1 ((exp(-t/tau_1) - 1) tau_1/t + 1)) +
69 (1-alpha) (tau_2 ((exp(-t/tau_2) - 1) tau_2/t + 1)))),
70 where T is the total time. alpha, tau_1 and tau_2 are obtained by fit‐
72 ting f^2(t) to error^2. When the actual block average is very close to
73 the analytical curve, the error is sigma``*``sqrt(2/T (a tau_1 + (1-a)
74 tau_2)). The complete derivation is given in B. Hess, J. Chem. Phys.
75 116:209-217, 2002.
76 Option -filter prints the RMS high-frequency fluctuation of each set
78 and over all sets with respect to a filtered average. The filter is
79 proportional to cos(pi t/len) where t goes from -len/2 to len/2. len is
80 supplied with the option -filter. This filter reduces oscillations
81 with period len/2 and len by a factor of 0.79 and 0.33 respectively.
82 Option -g fits the data to the function given with option -fitfn.
84 Option -power fits the data to b t^a, which is accomplished by fitting
86 to a t + b on log-log scale. All points after the first zero or with a
87 negative value are ignored.
88 Option -luzar performs a Luzar & Chandler kinetics analysis on output
90 from gmx hbond. The input file can be taken directly from gmx hbond
91 -ac, and then the same result should be produced.
92 Option -fitfn performs curve fitting to a number of different curves
94 that make sense in the context of molecular dynamics, mainly exponen‐
95 tial curves. More information is in the manual. To check the output of
96 the fitting procedure the option -fitted will print both the original
97 data and the fitted function to a new data file. The fitting parameters
98 are stored as comment in the output file.
99
101 Options to specify input files:
102 -f [<.xvg>] (graph.xvg)
104 xvgr/xmgr file
105 Options to specify output files:
107 -ac [<.xvg>] (autocorr.xvg) (Optional)
109 xvgr/xmgr file
110 -msd [<.xvg>] (msd.xvg) (Optional)
112 xvgr/xmgr file
113 -cc [<.xvg>] (coscont.xvg) (Optional)
115 xvgr/xmgr file
116 -dist [<.xvg>] (distr.xvg) (Optional)
118 xvgr/xmgr file
119 -av [<.xvg>] (average.xvg) (Optional)
121 xvgr/xmgr file
122 -ee [<.xvg>] (errest.xvg) (Optional)
124 xvgr/xmgr file
125 -fitted [<.xvg>] (fitted.xvg) (Optional)
127 xvgr/xmgr file
128 -g [<.log>] (fitlog.log) (Optional)
130 Log file
131 Other options:
133 -[no]w (no)
135 View output .xvg, .xpm, .eps and .pdb files
136 -xvg <enum> (xmgrace)
138 xvg plot formatting: xmgrace, xmgr, none
139 -[no]time (yes)
141 Expect a time in the input
142 -b <real> (-1)
144 First time to read from set
145 -e <real> (-1)
147 Last time to read from set
148 -n <int> (1)
150 Read this number of sets separated by &
151 -[no]d (no)
153 Use the derivative
154 -bw <real> (0.1)
156 Binwidth for the distribution
157 -errbar <enum> (none)
159 Error bars for -av: none, stddev, error, 90
160 -[no]integrate (no)
162 Integrate data function(s) numerically using trapezium rule
163 -aver_start <real> (0)
165 Start averaging the integral from here
166 -[no]xydy (no)
168 Interpret second data set as error in the y values for integrat‐
169 ing
170 -[no]regression (no)
172 Perform a linear regression analysis on the data. If -xydy is
173 set a second set will be interpreted as the error bar in the Y
174 value. Otherwise, if multiple data sets are present a multilin‐
175 ear regression will be performed yielding the constant A that
176 minimize chi^2 = (y - A_0 x_0 - A_1 x_1 - … - A_N x_N)^2 where
177 now Y is the first data set in the input file and x_i the oth‐
178 ers. Do read the information at the option -time.
179 -[no]luzar (no)
181 Do a Luzar and Chandler analysis on a correlation function and
182 related as produced by gmx hbond. When in addition the -xydy
183 flag is given the second and fourth column will be interpreted
184 as errors in c(t) and n(t).
185 -temp <real> (298.15)
187 Temperature for the Luzar hydrogen bonding kinetics analysis (K)
188 -fitstart <real> (1)
190 Time (ps) from which to start fitting the correlation functions
191 in order to obtain the forward and backward rate constants for
192 HB breaking and formation
193 -fitend <real> (60)
195 Time (ps) where to stop fitting the correlation functions in
196 order to obtain the forward and backward rate constants for HB
197 breaking and formation. Only with -gem
198 -filter <real> (0)
200 Print the high-frequency fluctuation after filtering with a
201 cosine filter of this length
202 -[no]power (no)
204 Fit data to: b t^a
205 -[no]subav (yes)
207 Subtract the average before autocorrelating
208 -[no]oneacf (no)
210 Calculate one ACF over all sets
211 -acflen <int> (-1)
213 Length of the ACF, default is half the number of frames
214 -[no]normalize (yes)
216 Normalize ACF
217 -P <enum> (0)
219 Order of Legendre polynomial for ACF (0 indicates none): 0, 1,
220 2, 3
221 -fitfn <enum> (none)
223 Fit function: none, exp, aexp, exp_exp, exp5, exp7, exp9
224 -beginfit <real> (0)
226 Time where to begin the exponential fit of the correlation func‐
227 tion
228 -endfit <real> (-1)
230 Time where to end the exponential fit of the correlation func‐
231 tion, -1 is until the end
232
234 gmx(1)
235 More information about GROMACS is available at <‐
237 http://www.gromacs.org/>.
238
240 2020, GROMACS development team
2412019.6 Feb 28, 2020 GMX-ANALYZE(1)