September 19, 2013, HFH 4164

Gene expression studies in the brain gave rise to data for thousands of genes, following two approaches: 1) gene-based (in the entire brain, as in the Allen Atlas),  2) cell-based (in a particular cell type). I propose a model to combine the power of the gene-based approach and the specificity of the cell-based approach, which can only be done computationally. The Allen Atlas of the adult mouse brain is used to estimate the region-specificity of 64 cell types whose transcriptional profile in the mouse brain has been measured in microarray experiments. The relevant optimization techniques have been implemented in the Brain Gene Expression Analysis toolbox (http://brainarchitecture.org/allen-atlas-brain-toolbox).

Pascal Grange obtained his PhD in theoretical physics from Ecole Polytechnique (Paris) in 2005. He worked on flux compactifications of string theory, generalized geometry and mirror symmetry, and stochastic  modelling. After a postdoc at
the IAS (Princeton) and Hamburg University, he evolved towards a more data-driven style of research. Over the last three years he worked at CSHL on the collective properties of gene expression maps in the brain, revealing anatomical properties of autism-related genes and region-specificity of cell types.