Research in Computational Chemistry
Docking: Computer scienctists and computational chemists join forces to develop highly parallelized docking procedures. This requires the optimization of the docking protocol. Different issues are tackled, as for example the treatment of structural water molecules, a problem most likely to occur when doing massiv molecular docking. Intelligent Pharma manages this problem by applying heuristics to determine how water molecules may be oriented in a regular docking experiment.
- O. Villacañas, S. Madurga, E. Giralt and I. Belda. Explicit Treatment of Water Molecules in Protein-Ligand Docking. Current Computer - Aided Drug Design. 5(3): 145-154 (2010)
De Novo Drug Design: Intelligent Pharma has recently marketed a computational tool called HELIOS, which uses a novel and powerful ligand-based virtual screening approach. Future goals of the computational chemistry group are HELIOS functionality for macromolecules and in designing drugs from scratch.
Actually, Intelligent Pharma's scientists already gathered a vast amount of experience towards achieving the mentioned objectives. Research towards using genetic algorithms for de novo peptide design on specified protein receptor surfaces has already been published. This is implemented in the tool known as ENPDA, which only requires the specification of the protein patch by the user to design different peptides able to bind to the target. Therefore, an evolutionary computation optimization kernel is used. ENPDA can also take advantage of previous information as the knowledge of a scaffold capable of binding to the target. This process is referred to as virtual combinatorial chemistry while the ab initio version (no previous information) is called de novo drug design.
I.
Belda, X. Llorà and E. Giralt. Evolutionary algorithms and de novo
peptide design. Soft Comput. 10(4): 295-304 (2006)- I. Belda, S. Madurga, X. Llorà, M.
Martinell, T. Tarragó, M. G. Piqueras, E. Nicolá and E. Giralt. ENPDA:
an evolutionary structure-based de novo peptide design algorithm. J.
Comput. Aided. Mol. Des. 19(8): 585-601 (2005)
HELIOS 2.0 will use the same computational tools as ENPDA. For now, they are optimized towards designing small molecules from scratch by using the structure of any previously defined ligand.
Flexible QSAR: Classical QSAR has typically shown high error rates in drug design projects. At Intelligent Pharma we have identified possible reasons for the occurance of errors, which are proposed to be solved by several mathematical innovations.
In classical QSAR projects, molecular descriptors are fixed values for each molecule of a training set, i.e. molecular weight, isoelectrical point, etc. New functional molecular descriptors being developed by Intelligent Pharma are not constant values, but functions that depend on the internal flexibility of the molecules. Therefore, a molecular descriptor becomes functional as, for instance, a function is used to describe the volume of a molecule due to its degrees of internal flexibility. For the determination of functional data different tools as for example customized machine learning are applied.