Abstract

Advances in biological understanding are increasingly reliant on quantitative modeling and theory that capture critical aspects of biological phenomena. In developmental biology and in the study of cellular dynamics, quantitative data are available on the activity of genes in individual cells or cell-type samples.  But how should we formalize the intuitive concept of cell-type specificity? This talk presents measurements of cell-type specificity based on the amount of information that a gene?s level of expression provides about the identity of the cell. The approach is applied to global cell-type specific data available in the Arabidopisis root and the mouse brain, revealing the modular structure of multicellular transcriptomes involved in tissue patterning.  Applications to biomarker discovery and treatment-specific data sets will also be discussed.