This document discusses using a chemical interaction matrix and biclustering approach to analyze multi-dimensional biochemical and chemical biology data to better understand mechanisms of action. The procedure involves assembling a matrix of compounds versus activity and descriptive features, normalizing the data, folding activity measurements into the features, and then using biclustering to identify contiguous regions within the matrix that may indicate groups of compounds sharing a mechanism of action or features correlated with activity. An example application to a dataset of 773 molecules and 298 molecular descriptors relating to oral bioavailability achieves promising results, with biclustering condensing the data into 18 high-quality clusters that cover most of the training space and should provide a strong basis for target identification when compared