We are pleased to announce the release of the first open source version of the python module gaddlemaps, which implements the GADDLE Maps algorithm. This algorithm is able to map molecular systems between two different resolutions (for example between coarse grained and fully atomistic resolutions) without any user intervention.
The algorithm does not rely on precalculated configurations, instead it uses an optimisation algorithm to find the best mapping between the species in the two resolutions. This process is performed for only one molecule of each specie and it extrapolate these results to all the other molecules of the system. This makes the algorithm very quick, being able to map systems with thousand of simple molecules in matter of seconds. It also includes optimisations for big proteins, for which the algorithm does not usually take more than a couple of minutes.
The system is designed to work with coordinate and topology files from gromacs (.gro and .itp) and should work with files from any Gromacs version up-to and including Gromacs 2020.
The python module includes 2 version of the optimisation algorithm, one that is built purely in python, making it very portable, and a c++ compiled version that is much faster. The module includes also a command line interface to perform simple maps and some examples of how to use all the available options for the mapping (including a fully interactive Jupyter notebook ).
The module can be easily installed from Pypi using “pip install gaddlemaps” and it is hosted in a GitHub repository (GitHub - txemaotero/gaddlemaps) where you can find a README with all the installation instructions, the usage of the command line interface and some FAQs. In the GitHub repository you can also find the code-tests (with over a 90% of coverage) and all the examples.
Moreover, you can also access to the documentation of the python module in readthedocs.
The algorithm behind the gaddlemaps module has been published and the code has been tested for several systems, from simple molecular solvents to proteins.
Support for the library is performed in the GitHub page by opening new issues. Moreover you may contact us by email: