Datamining of the Allen Atlas
The amount of genetic data generated is increasing at enormous speed. We need strategies to analyze and extract usable information from that data. One possibility is to use the averaged expression pattern of groups of genes to derive anatomical hepothesis from genetic screens. For example, in GWAS many genes are above significance but not when multiple comparison corrections are used. We postulate that by grouping genes according to clusters of coexpression we may be able to generate anatomical hypotheses about the phenotype studied in the GWAS. To do that, we wrote a phython script that given a list of genes goes to the Allen Atlas database, extracts the normalized volumes and averages the expression of all those genes into one volume.
As an example, we used a list of genes involved in nicotine addiction, from Knowledgebase for Addiction Related Genes (KARG) , and found some expected areas such as the striatum, the cortex and the habenula.