An approach toward the problem of outliers in QSAR.

Compounds that have unexpected biological activity and are unable to fit in a QSAR model are known as outliers. These are valuable in defining the limitations under which compounds act by a common molecular mechanism modeled by one or more descriptors, and also in defining the experimental limitations of the biological test data. Thus, the outliers should be submitted to particular attention to see if the reason for their peculiarity can be determined. Separating these outliers from the main data set and formulating another QSAR can resolve the problem. Our result shows that these outliers may be acting by a different mechanism or interacting with the receptor in different modes.