Abstract
As in several other psychiatric disorders, genome wide association studies have not revealed loci involved in anorexia nervosa (AN). Most likely, the comparably small number of DNA samples represents the major limitation; at least one genome wide significant finding is the rule if 10,000 patient samples are included. However for AN, the two GWAS conducted up to today have in total included less than 5,000 patients; a joint analysis of the data sets has not yet been performed. Based on these two GWAS we can currently conclude that common variants with strong effect sizes are unlikely to account for the predisposition to AN in a substantial subgroup of patients. Complex disorders are due to a multitude of factors, clearly requiring large sample sizes. This entails the need to discuss future biomedical research venues for AN making use of cutting-edge technology. We in parallel need to increase sample sizes. We propose to focus on clinical research addressing diagnostic, therapeutic and prognostic aspects. Meta-analyses should be performed for diverse questions. In order to perform these, we need a consensus as to what questions require what data, so that these can be collected and analyzed by individual groups and at the same time feed a data base that allows interested researchers to perform meta-analyses. We need to realize that most of our biomedical research into AN pertains to state and not to trait markers; biomarkers are related to AN associated starvation and not to the mechanisms underlying the development of AN. An open discussion on biomedical research is not only important to focus our scientific goals but also to attract young scientists into our field.

Keywords: Anorexia nervosa, GWA (genome-wide association study).