Designing better drugs

Figure 1.14: On the left, several familiar drugs are pictured as ball-and-stick models (courtesy of the New York University MathMol Library [734]). On the right, 3D models of protein-ligand docking are shown from the AutoDock software package (courtesy of the Scripps Research Institute).
\begin{figure}\begin{center}
\begin{tabular}{ccc}
\psfig{file=figs/caffeine.ps,h...
...truein} \\
Nicotine & THC & AutoDock \\
\end{tabular}\end{center}
\end{figure}

Planning algorithms are even impacting fields as far away from robotics as computational biology. Two major problems are protein folding and drug design. In both cases, scientists attempt to explain behaviors in organisms by the way large organic molecules interact. Such molecules are generally flexible. Drug molecules are small (see Figure 1.14), and proteins usually have thousands of atoms. The docking problem involves determining whether a flexible molecule can insert itself into a protein cavity, as shown in Figure 1.14, while satisfying other constraints, such as maintaining low energy. Once geometric models are applied to molecules, the problem looks very similar to the assembly problem in Figure 1.3 and can be solved by motion planning algorithms. See Section 7.5 and the literature at the end of Chapter 7.

Steven M LaValle 2020-08-14