Initialization: define BCUT chemistry space and specify a distance cutoff value, c based on Distance Threshold calculated below. Iteration: for each compound, j, in the candidate compound collection, Calculate its distance to the nearest neighbor from the current compound collection, S: $$D_j=\underset{i}{\text{min}}\left|y_j-x_i\right|$$ yj is the descriptor vector of candidate compound j, and xi is the descriptor vector of compound i in the current compound collection, S. If the distance to the nearest neighbor Dj > c, then add the compound j into the current compound set: S ← S + candidate compound j. Go to step 2 to analyze next candidate compound.