To assess the utility and accessibility of computational tools currently available for molecular designers, a representative list of available computational tools was generated using www.click2drug.org, an online “directory of computer-aided drug design tools” maintained by the Swiss institute of bioinfor- matics.52 the list includes single-function QSARs and collections of ADMET prediction models, with a mix of downloadable software, databases, and web-based platforms. the list includes the tools that the authors considered most useful and accommodating to general trends for the overlapping fields of computational toxicology, computational chemistry, and computational medicinal chemistry.
ACD Percepta (www.acdlabs.com)
this medicinal chemistry software42 is distributed by Advanced Chemistry Development.53 the program accepts single and batch .sdf or .mol files and also allows users to draw structures directly in the software. the outputs include basic physicochemical parameters (logp, H-donors/acceptors, rotational bonds, rings, Lipinski's rules violations, and solubility) as well as ADME (Caco-2 permeability, plasma protein binding, and CNS penetration), and some drug safety information (some cytochrome p450 (CYp) inhibition tests, Ames, and hERG). the program provides color-coded interpretations of those outputs and indicates on sliding scales how drug-like a molecule is. navigation is easy, and different compounds can be compared in convenient tables or tabular format. Saving and sharing data is simple through export into either .pdf or .csv format, and the user interface is aesthetically pleasing and easy to work with. Of particular utility is the Structure Design Engine module, as it allows the chemist to edit a molecule in a drawing window and observe how structural modifications affect predicted physicochemical and ADME properties in real time: as the molecule is modified, sliders move and alerts appear based on structural changes. Additionally, this module is capable of proposing a set of analogs for a given molecule based on a set of desired physicochemical parameters—a feature which greatly simplifies the “iteration” phase of the workflows. It is worth noting that this (like most of the programs discussed herein) is intended for the development of pharmaceuticals, not industrial chemicals, so appropriate care must be taken when interpreting the results generated by ACD percepta, as the molecule of interest may fall outside the training sets used to develop the predictive algorithms. For chemists looking to improve the ADMET profile for a compound, this software offers most of what is needed.