Multiparameter lead optimization

Segall, M., Champness, E., Leeding, C, Lilien, R., Mettu, R., and Stevens, B. (2011) Applying medicinal chemistry transformations and multiparameter optimization to guide the search for high-quality leads and candidates. Journal of Chemical Information and Modding, 51, 2967-2976. 

The critical components in PEMFCs are the CLs, where the electrochemical reactions take place that transform oxygen molecules into water at the cathode and hydrogen molecules into protons at the anode. Optimization of the CLs is a multiparameter problem, since they must meet numerous requirements simultaneously, ensuring optimum rates of electrocatalytic reactions as well as providing pathways for electrons, protons, reagents, and products. The strongest constraints are imposed on the cathode, where water molecules produced electrochemically must be transferred to the gas phase and removed from the catalytic layer to avoid its flooding. The latter leads to a drastic decrease in cell performance. 

Furthermore, early models had to be simple because hand calculation was necessarily the mode, but now computer-based "process simulators" readily solve complex, multiparameter equations. Such simulators enable us to generate alternative "what-if" scenarios to study feasibility and optimization they also allow us to probe the smallest details of process facilities and conditions. However, this powerful capability is limited by the approximations we provide to the simulator and by our interpretations of the output that the simulator provides to us. Casual, uncritical use of process simulators can obscure the significance of results and lead to process designs that are physically unrealizable. Therefore, you must give some attention to the accuracy with which property values will be needed and to the computational resources that will be required to achieve the required accuracy. 

Another highly relevant area of research concerns the analysis of multiparameter SAR. In the course of a lead optimization project, the number of considered endpoints and parameters grows considerably. This makes it likely to observe tradeoffs between different parameters that cannot be optimized separately due to parallel SAR. Furthermore, one is often forced to work with incomplete or even sparse datasets— most compounds are characterized with respect to some endpoints, but a full profile is available only for a small fraction of the compound set. Methods that support the analysis of such complex optimization scenarios in a convenient, efficient, and intuitive way are highly welcome. 

---