Screen, inactive compounds

Although the programs for this type of application were initially developed for relatively small numbers of compounds, they have also been applied to sets of screening data. Software to identify common structural features associated with active compounds, yet not present in the inactive compounds, using 2D descriptions of the compounds has been developed (57,58). 

In addition to this pharmacophore hypothesis, although it met only three of the four criteria, model 1 from run 6 was retained. Surprisingly, despite criterion number 2 not being satisfied (RMS= 1.62, r=0.79), this model exhibits a remarkable ability to discriminate between active and inactive compounds as assessed by the ROC curve, AUC=0.95. In contrast, model 1 from run 8 has good statistics (RMS=0.76, r= 0.96) but a lower AUC of 0.87. This illustrates that a good model for activity prediction may not be the best for virtual screening applications. Let us analyze these two pharmacophore hypotheses further. 

Another report reveals, however, that in vitro DNA-binding assays are insufficient to predict platinum antitumor activity [15]. Primer extension (Fig. 1) was used to identify specific adducts formed by platinum complexes on DNA in HeLa cells. The DNA adduct profile correlated well with in vivo antitumor activity for cis- and trans-DDP, Pt(en)Cl2, and two acridine-tethered platinum complexes. When the complexes were allowed to react with purified DNA in solution, there were no substantial differences in adduct profiles between active and inactive compounds. This result demonstrates that cell-based assays can be better predictors of in vivo activity than in vitro assays, particularly when the in vitro screen does not require aunique, mechanism-based molecular interaction. 

Although it is essential to test promising compounds in mice and other animal models prior to human trials, it is economically, ethically and often scientifically preferable to use cell-based and in vitro approaches to eliminate inactive compounds before commencing animal trials. Clearly, animal models are not an appropriate screen for combinatorial libraries of platinum complexes they should be used to study further the promising leads identified by high-throughput methods. 

While the attributes of a drug-like profile are in no way to be minimized, the first and foremost properties for a proposed test series during early drug discovery is certainly their assay-likable profile - that is, if it cannot be tested you will not get a result, or even worse, one may perceive that one has an inactive compound and thus erroneously derive a negative SAR data point. Indeed, for the case of esmolol, the entire discovery process was almost halted when its first two series of target compounds were either too insoluble or too plagued by ISA, to be able to be effectively screened. 

Figure 2 Workflows for experimental physical HTS and virtual screening of compound libraries and their combination, for active and inactive componnds, absence and presence of activity are determined with the same degree of reliability. This reliability is an advantage for bnilding SAR models. In addition, the analysis of the dose response curve shapes allows some...

---