MD-jeep: a software tool for Distance Geometry

MD-jeep is a software tool for solving molecular distance geometry problems (MDGPs). It is able to solve a subclass of instances of the problem for which a discrete reformulation can be supplied. We employ a Branch & Prune (BP) algorithm for the solution of discretizable MDGPs.

MD-jeep is the result of a strong collaboration among Antonio Mucherino, Leo Liberti, Carlile Lavor and Nelson Maculan. Over the years, PhD and postdoc students under our supervision have also been contributing to this research topic. The new method for the computation of atomic coordinates in MD-jeep v.0.2 was developed in collaboration with Douglas Soares Gonçalves, who was postdoc student in Rennes for one year. Other experimental versions of MD-jeep (not available for download) have also been developed. Many details regarding the discretization of the problem and the BP algorithm can be found in our publications (the complete list of papers can be found here).

MD-jeep is distributed under the GNU General Public Licence (v.2). If you experience problems when using our software, please let us know!

Some Test problems are available for download. All test problems are related to proteins that exist in nature and whose three-dimensional conformations are available on the Protein Data Bank (PDB). The first set contains instances formed by the sequence of backbone atoms N, Cα and C. These atoms are extracted from PDB files and the distance between atom pairs is included in the instance if, and only if, this distance is shorter than 6 Angstroms. This is done for simulating data obtained through experiments of Nuclear Magnetic Resonance (NMR), that generally provide short range distances.

The second set contains instead instances that are more realistic. The entire protein backbone is considered and represented by bond lengths and angles, while distances between pairs of hydrogen atoms are considered only when the distance falls into the interval [2,5] Angstroms. In fact, NMR experiments are generally able to detect only distances between hydrogens. Finally, in the third set of instances, the side chains (belonging to each amino acid forming the protein) are also included.

Download: Test problems 1, Test problems 2, Test problems 3.

The current version of MD-jeep is able to work only with instances containing precise distances. The extension to interval data is under development (refer to our papers to follow the progress) and it will be available in future releases.

Back Home