Skip to content

HITS-MBM/conan

Repository files navigation

Requirements

In order to successfully run CONAN you need to have working installations of:

  • Python 3, with numpy and scipy installed. (Anaconda 3+ recommended). Using an older version of scipy can cause an error in testing. The error will read something like: ValueError: Valid methods when the raw observations are omitted are 'single', 'complete', 'weighted', and 'average'. If you are NOT interestred in cluster analysis, feel free to ignore this error in testing.
  • GROMACS molecular dynamics engine, version 5.0 or later. GROMACS can be downloaded from the official website. (NB: GROMACS is only used for the first step of analysis, the MD trajectory can be obtained by other programs too.)
  • Gnuplot plotting tool. It can be downloaded from the official website.
  • Mencoder is obtained through MPlayer, which can be found on the official website.

In fact, even gnuplot and mencoder are optional (but recommended!).

Usage

CONAN reads in a single input file and runs. For example, if the input file is input.inp, write:

./conan.py input.inp

And you are good to go. Note that, for a first run, CONAN will run gmx mdmat, which will require you to choose a group of atoms.

Memory requirements

Since the involved datasets scale, in principle, as N_(frames) x N_(res)^2, you should be mindful of disk space and memory. The main data file on the disk is dmf.xpm, which scales roughly as ~N_(frames) x N_(res)^2 x 2B. For example, a 100-frame simulation on a 100-residue protein will take up about 2MB of disk space.

The output of CONAN will be many .png files and also many .dat files, which contain the raw data for the former. The .dat files can take up a lot of space; about ~N_(frames) x N_(res)^2 x 20B, if all the raw matrices are to be stored. If this is too much, you can try the keyword CLEAN_MATRICES (creating .png files only) or MATRICES no and MAKE_MOVIE no (creating only aggregate plots).

The memory requirements of CONAN are usually not too extreme, since data storage is done in a sparse way. Still, if you find that this does not help, consider using the "economy mode" (keyword ECONOMY), which does only basic statistics on the data.

In general, it is a good idea to run CONAN on a subset of your data (setting DT high, or restricting BEGIN or END to a part of the simulation) before running it on the entire set.

Testing CONAN

You can test your CONAN distribution by:

cd test/
./test.py

And read the results on the screen. The test script cleans up after itself, unless there are errors. Please note the previous remark -- older versions of Scipy are known to return a ValueError on cluster analysis.

CONAN tutorials

There are some sample applications (input files and trajectories) in the folder applications/. Check out our tutorials on https://contactmaps.blogspot.de/ for step-by-step instructions.

Also check out other example input files that you can use on your own trajectories at https://github.com/HITS-MBM/conan/blob/master/docs/Example_inputs.md

Using parts of CONAN

You can use CONAN as a library by import conan in a Python script. The variable names are relatively easy to understand, although not everything is 100% user-friendly at the moment.

Where to reach us?

You can contact me (Csaba) at csaba.daday at h-its.org. Check out the blog at https://contactmaps.blogspot.de/ and the YouTube channel at https://tinyurl.com/Con4nMD .