Flow cytometry is a common technique for quantitatively measuring the expression of individual molecules on cells. The molecular expression is represented by a frequency histogram of fluorescence intensity. For flow cytometry to be used as a knowledge discovery tool to identify unknown molecules, histogram comparison is a major limitation. Many traditional comparison methods do not provide adequate assessment of histogram similarity and molecular relatedness. We have explored a new approach applying information theory to histogram comparison, and tested it with histograms from 14 antibodies over 3 cell types. The information theory approach was able to improve over traditional methods by recognizing various non-random correlations between histograms in addition to similarity and providing a quantitative assessment of similarity beyond hypothesis testing of identity.