Information-theoretic biodescriptors for proteomics maps: Development and applications in predictive toxicology

This paper describes an approach using information theory to derive a new complexity measure for proteomics maps generated using 2-dimensional gel electrophoresis. The maps used in this study were partitioned into 5×5 grids and the total abundance of protein material in each grid was compared to the total abundance for the entire map. Next, Shannon's relation was applied to characterize the distribution of spots across the proteomics map. Details of the approach are discussed here, including an illustrative example and an example of the calculations for a proteomics map containing 200 spots. Finally, results for the Map Information Content index are presented for a set of five maps calculated using 200 spots, 500 spots, and 1,054 spots. It is hoped that the application of information-theoretic techniques to characterize the complexity of these maps, thus reducing the amount of information presented to the researcher, will help in the analysis and comparison of maps containing a great deal of information.