“In vivo veritas” is the motto of the pharmacologist, but we usually have much more information about the in vitro effects of a developmental drug. DPDMODEL attempts to predict in vivo drug efficacy and selectivity by first analyzing in vitro data from two or more cell lines (for example from a synthetic lethal screen), and second by obtaining pharmacokinetic (PK) data for that drug in the species of interest. It then integrates the selective pharmacodynamic effect over the PK time course.
This approach has been applied to prediction of drug interactions (synergism and antagonism) from in vitro data. A drug combination may show synergism against tumour cells in vitro, but this will not translate into a therapeutic advantage if that combination is equally (or more) synergistic against normal cells. Conversely a combination may be antagonistic, but can still be useful if it shows enhanced selectivity.
The ideal drug combination will be synergistic against tumour cells and antagonistic against normal cells. Such combinations have been shown to exist and DPDMODEL provides strategies for searching for them.