Quick Start¶
A basic work-flow with influx_si
is composed of the following steps:
- Create a FTBL file describing your metabolic reactions, carbon transitions, experimental data and some options. Let call an example file
mynetwork.ftbl
. The FTBL file must follow syntax rules elaborated for 13CFlux software. The FTBL file is a plain text file. The syntax rules will be more or less obvious for someone working on metabolism biochemistry. So, to go quickly, you can inspire from an example filetest/e_coli.ftbl
distributed with theinflux_si
software.
Note
Staring from the version 2.5,
NA
values (as “Non Available”) are admitted as measurements values where appropriate. The difference with FTBL where they are simply omitted is that NA measurements are simulated and are present in the vectorssimulated unscaled labeling measurements
andsimulated scaled labeling measurements
in the result kvh file.Note
In case of
influx_i
, label kinetics can be provided in a separate plain text file with values separated by tabulations. First column in this file gives measurement names, and all other columns correspond to a particular time point each. Time points are given on the first line of the file. In this file, there can be more time points than were used in a real experiment for sample harvesting. In this case, the labeling is simulated and reported for these fictitious time points but the least squares fitting is obviously done only at points where real data are reported.Empty cells in this file are equivalent to
NA
. Note also that is _not_ necessary to introduce empty columns at regular intervals just to increase the time resolution precision. There is a parameternsubdiv_dt
that is designed for this purpose. If it is greater than 1, each time interval defined in the text file is divided innsubdiv_dt
sub-intervals.
Set your current directory to the directory of
mynetwork.ftbl
and run$ influx_s.py mynetwork
or:
$ influx_i.py mynetwork
Depending on stationary or instationary labeling context. Note that the suffix .ftbl
is optional and influx_si
installation directory is supposed to be on the PATH.
The
influx_si
run will produce the following files in the same directory thatmynetwok.ftbl
mynetwork.log
- containing the run-time output from various scripts, in particular, it contains a report on convergence history during the fitting process. It can be helpful for identifying potential problems but if everything is going well, the user does not have to examine the content of this file;
mynetwork.err
- containing the warning and error messages. Normally, this file should be empty (0 byte size);
mynetwork_res.kvh
- containing all of the results. KVH format is a lightweight plain text format for hierarchically structured data. It can be seen in a text editor or in a spreadsheet software as its fields are tab separated. It can also be processed by user’s custom software for post-processing, graphics output and alike. If
influx_si
is run on a series of starting points there will be generated a common result filemynetwork_res.kvh
containing common information to all starting points but also a series of kvh files, one by starting point, e.g.mynetwork_res.V1.kvh
,mynetwork_res.V2.kvh
and so on;mynetwork.pres.txt
- containing a matrix of fitted parameters and final cost values. Each column corresponds to a particular starting point if run with
--fseries
and /or--iseries
options. Ifinflux_si
was run without these options, the file will contain only one column corresponding to the starting point defined in themynetwork.ftbl
file.edge.netflux.mynetwok
,edge.xchflux.mynetwok
,node.log2pool.mynetwork
as the middle name of this files suggest, they can be used to map the corresponding values on the network graph in the cytoscape software.
Note
All these files are silently overwritten if already exist. So take care to copy your results elsewhere if you want to protect them from overwriting.
custom files (e.g.
mynetwork.pdf
) These files can be produced by user supplied scripts that are executed at the end ofinflux_si
simulations. For example, we provide a scriptplot_imass.R
which can be used to plot label kinetics obtained byinflux_i
. One or many of such custom scripts can be given in FTBL file, section OPTIONS, fieldposttreat_R
(cf. e_coli_i.ftbl for example)Note
It can be helpful to do some “dry runs” by executing
$ influx_s.py --noopt mynetworkbefore collecting actual measurement data to see if intended measurements will be sufficient to well define all fluxes or at least the fluxes of interest. It is possible to do because the measurement values in the FTBL file does not matter for flux SD calculation when
--noopt
option is used. So it can be used any values even NA at this moment. In the contrary,dev
values set in measurement sections of the FTBL file, must be realistic. It is generally not a problem as they express measurements errors and are more or less known for a given measurement chain.It is worthwhile to stress that a “dry run” is done for some presumed free fluxe values and if they reveal to be very different from actual flux values, it can happen that a network considered as well defined at moment of “dry run” turned into a badly defined network with actual measurement data and corresponding estimated fluxes. So it is important to do his best to guess the most realistic free fluxes for “dry runs”.
- See warning and error messages in
mynetwork.err
if any. Correct what has to be corrected and retry p. 2 - Extract and use the numerical results from the
mynetwork_res.kvh
file. - Optionally, visualize net fluxes (or exchange fluxes or logarithm of metabolite concentrations \(\log_2(M)\)) in cytoscape using
edge.netflux.mynetwok.attrs
,edge.xchflux.mynetwok.attrs
ornode.log2pool.mynetwork.attrs
.