Gracias a la colaboración de muchas personas, el proyecto

"Busqueda de Medicamentos para la Leishmaniasis"

ha recibido casi 19 millones de resultados.

Del billón de resultados procesados por el supercomputador del World Community Grid de IBM, casi 19 millones corresponden al proyecto "Busqueda de Medicamentos para Leishmaniasis" y que se esta realizando en la Universidad de Antioquia en Medellin-Colombia. Esto ha sido posible por la colaboracion altruista de muchas personas que entienden el valor de donar el tiempo de ocio de sus computadores para apoyar una causa que puedan beneficiar mas adelante a comunidades afectadas, en nuestro caso, por Leishmaniasis. Muchas Gracias a todos por apoyar esta iniciativa.

Drug Search for Leishmaniasis Update

Rodrigo Ochoa and Carlos Muskus

The project “Drug Search for Leishmaniasis” has been running during 10 months, and the following is a brief description of our progress and what have been validated until now:

The original project included the evaluation of 53 crystallized Leishmania proteins obtained from the PDB database along with 5,300 modelled structures, (100 modelled structures for each protein downloaded from PDB) (figure 1). The modelling was carried out through Molecular Dynamics using the NAMD program, representing an indirect way to provide flexibility for each one of the potential targets.

However, based on the statistic of the project after 6 months of activity, to run the 5.353 protein structures against the 600.000 compounds will spend around 20-25 years to complete. These made us to reconsider the project just prioritizing 10 structures for each of the 53 proteins for a total of 530 structures (Phase 2). Originally the idea was to build a molecular trajectory using snapshots of a protein during a computational simulation, trying to mimic the biological environment of the protein inside the organism. The trajectories contain 100 snapshots. Nevertheless, to prioritize the data, 10 snapshots per trajectory (i.e from the 100 structures per protein, the snapshots 10, 20, 30, 40, 50, 60, 70, 80 90 and 100 were selected) (see figure 2). That rational selection pretends to cover substantial changes in the conformation of the binding site of the protein.

Currently, WCG have processed approx. 76.200.000 docking simulations (127 structures against 600.000 compounds in phase I). From the second phase, it has processed approx. 37.800.000 docking simulations (63 priority structures against the same 600.000 compounds). Due to the huge amount of data, only the top 20 compounds per structure (compounds with the best score hits) and based on the score function proposed by the AutoDock VINA software, have been extracted. As an example, we provided the top 20 compound which lists from the protein 1-10, 1-20 and 1-30* (See tables 1, 2 and 3)

*The notation 1-10 means the modelled structure 10, from the protein target 1 of the 53 original proteins downloaded from PDB database.

A future analysis with the high relevant results will be published in the next update of the project.

We want also to thank all the WCG members supporting this project all over the world, for the almost 19.000.000 results processed, out of the 1 billion data returned to the World Community Grid from IBM.

Figure 1. This new graph represent the current progress of the 583 priority targets (530 modelled proteins + 53 original structures) docked against the 600.000 compounds. The orange bar at the bottom of the figure represents the 53 original proteins structures already docked. The others bars represent the modelled protein structures (from 10 to 100). The blue color areas represent modelled proteins already docked and the green color areas represent the modeled proteins pending for docking. As you can see, there are 136 modelled structures still pending for docking which correspond to approximately 29%.

Figure 2. Global progress of the project. Up to date (April 10) 71% of the project have been completed.