Pharma Algorithms

Pharma Algorithms
Reviewed by Derek Reynolds

The ADME Box software is a convenient desktop application that runs under Windows XP or Windows 2000. It provides a set of databases and prediction tools for estimating the physicochemical properties and the oral bioavailability of molecules. ADME Boxes can be configured as a complete package or users can purchase a selection of the available modules. The modules appear as side tabs and clicking on the tabs switches between modules. In each module there is a structure pane that displays a 2D representation of the chemical structure used for calculation. Structures can be pasted in from another application, such as ISIS Draw or ChemDraw, or input as a structure file or a smiles string. Immediately after loading a new molecule the calculation is performed automatically and a few seconds later the results are displayed. All single structure calculations are also available in batch mode from the batch calculations module.

The ionisation module estimates the total number of ionisable groups and predicts the principal pKa values. Per cent fractions of different ionic forms are displayed graphically as is the calculation and display of octanol/water logD as a function of pH. Solubilities in aqueous buffer can also be estimated and displayed graphically versus pH. The QSAR algorithm for calculating solubility was developed using a training set of about 6000 compounds and the experimental data for these compounds is available.

The database interface is simple to use and can be searched using numerical, text and substructure queries. The same generic database interface is used for other modules and includes experimental values for bioavailablity (700 drugs), human oral absorption (800 drugs) and interactions with the P-glycoprotein efflux system (2000 drugs). For a novel structure the oral bioavailability module provides overall estimates of probability is based on predictions of solubility, stability, permeability, interactions with transporter proteins, and susceptibility to first pass metabolism. Each of these factors is predicted using independent algorithms developed using specific datasets.

Overall the software is easy to navigate and should be useful to scientists involved in drug discovery both to help them interpret the results of in-vivo experiments and to help design new molecules with improved ADME characteristics.