Selection of molecules with many attributes may be achieved using a combination of
principal components analysis
(PCA) and
design of experiments.
Chemical maps produced by this method enable chemists to visualise the properties of their compounds in a rapid and intuitive manner. Powerful features such as the new Onion Design family ensure representative choices despite thousands of descriptor variables. Partial least squares projections (PLS) is also an ideal technique to optimise desired versus undesired activity.
For more detailed information, see the application notes:
"Data Mining", "Statistical Molecular Design", and “Chemical Maps using SIMCA-P”.
