The Tailored Approach to Molecular Similarity

Finally, the third set of indices was selected based on a ridge regression model using all 267 indices to model mutagenicity. From the modeling results, 111 -values were extracted and the 20 top indices were selected as axes for developing the similarity space. A summary of the correlation coefficients for 

As can be easily observed from the table, tailoring the selected set of indices significantly improved the estimative power of the model, resulting in roughly a 10% increase to the correlation coefficient. These results, as with all of the results we have seen from tailored similarity spaces, are promising and we beheve that tailored spaces will be very useful both in drug discovery and toxicological research. 

---