A custom collaboration with: Quintiles
To develop better therapeutics and to do so more efficiently, pharmaceutical scientists need advanced approaches to predicting the potential effects of a specific compound. “The history of drug discovery tells you that simple analysis based on twodimensional chemical structure is of limited value,” explains Steve Gardner, chief operating officer at Cresset in Hertfordshire, UK. “Lots of compounds that look similar in structure have very different properties, and very different structures can have very similar activity.”
Pharmaceutical discovery will eventually push past that problem with multi-dimensional in silico techniques using computing, modeling and simulation. Some of those methods already exist. For instance, Cresset has developed software that models the interactions of a compound by using four property fields: positive electrostatics, negative electrostatics, van der Waals forces and hydrophobicity. These fields can be generated for a potential therapeutic and its target. The software then identifies extrema in the fields, which are referred to as field points. “These field points are the likely points of interaction between the compound and the protein target,” says Gardner.
Although such approaches are not entirely new, they keep improving. For example, Gardner points out that many of the existing techniques for determining the fields around a compound were designed in the 1980s, when less computing power forced researchers to take more shortcuts, such as assuming atom-centered charges. “You can pretty much guarantee that the center of an atom is the one place that the charge will not be,” says Gardner. Cresset’s proprietary approach puts the charge where it really is, and that makes more accurate fields.