Often, one would want to generate smooth data from a fit over a small number of data points. For an example, take the data in the
following file. It contains (fake) experimental data points that obey to Michaelis-Menten kinetics:
$$v = \frac{v_m}{1 + K_m/s}$$
in which \(v\) is the measured rate (the y values of the data), \(s\) the concentration of substrate (the x values of the data), \(v_m\) the maximal rate and \(K_m\) the Michaelis constant. To fit this equation to the data, just use the
fit-arb
fit:
QSoas> l michaelis.dat
QSoas> fit-arb vm/(1+km/x)
After running the fit, the window should look like this:
Now, with the fit, we have reasonable values for \(v_m\) (
vm
) and \(K_m\) (
km
). But, for publication, one would want to generate "smooth" curve going through the lines... Saving the curve from "Data.../Save all" doesn't help, since the data has as many points as the original data and looks very "jaggy" (like on the screenshot above)... So one needs a curve with more data points.
Maybe the most natural solution is simply to use
generate-buffer
together with
apply-formula
using the formula and the values of
km
and
vm
obtained from the fit, like:
QSoas> generate-buffer 0 20
QSoas> apply-formula y=3.51742/(1+3.69767/x)
By default,
generate-buffer
generate 1000 evenly spaced x values, but you can change their number using the
/samples
option. The two above commands can be combined to just one call to
generate-buffer
:
QSoas> generate-buffer 0 20 3.51742/(1+3.69767/x)
This works, but it is quite cumbersome and it is not going to work well for complex formulas or the results of differential equations or kinetic systems...
This is why to each
fit-
command corresponds a
sim-
command that computes the result of the fit using a "saved parameters" file (here,
michaelis.params
, but you can also save it yourself) and buffers as "models" for X values:
QSoas> generate-buffer 0 20
QSoas> sim-arb vm/(1+km/x) michaelis.params 0
This strategy works with every single fit ! As an added benefit, you even get the fit parameters as meta-data, which are displayed by the
show
command:
QSoas> show 0
Dataset generated_fit_arb.dat: 2 cols, 1000 rows, 1 segments, #0
Flags:
Meta-data: commands = sim-arb vm/(1+km/x) michaelis.params 0 fit = arb (formula: vm/(1+km/x)) km = 3.69767
vm = 3.5174
They also get saved as comments if you save the data.
Important note: the
sim-arb
command will be available only in the 3.0 release, although you can already enjoy it if you use the
github version.
About QSoas
QSoas is a powerful open source data analysis program that focuses on flexibility and powerful fitting capacities. It is released under the
GNU General Public License. It is described in
Fourmond, Anal. Chem., 2016, 88 (10), pp 5050–5052. Current version is
2.2. You can download its source code and compile it yourself or buy precompiled versions for MacOS and Windows
there.
No comments:
Post a Comment