Nat Genet. 2010 Aug 15. [Epub ahead of print]
Hor H, Kutalik Z, Dauvilliers Y, Valsesia A, Lammers GJ, Donjacour CE, Iranzo A, Santamaria J, Adrados RP, Vicario JL, Overeem S, Arnulf I, Theodorou I, Jennum P, Knudsen S, Bassetti C, Mathis J,Lecendreux M, Mayer G, Geisler P, Benetó A, Petit B, Pfister C, Bürki JV, Didelot G, Billiard M, Ercilla G, Verduijn W, Claas FH, Vollenwider P, Waeber G, Waterworth DM, Mooser V, Heinzer R, Beckmann JS, Bergmann S, Tafti M.
 Center for Integrative Genomics, University of Lausanne, Lausanne, Switzerland.  Institut National de la Santé et de la Recherche Médicale U888, Montpellier, France.  These authors contributed equally to this work.
Narcolepsy is a rare sleep disorder with the strongest human leukocyte antigen (HLA) association ever reported. Since the associated HLA-DRB1*1501-DQB1*0602 haplotype is common in the general population (15-25%), it has been suggested that it is almost necessary but not sufficient for developing narcolepsy. To further define the genetic basis of narcolepsy risk, we performed a genome-wide association study (GWAS) in 562 European individuals with narcolepsy (cases) and 702 ethnically matched controls, with independent replication in 370 cases and 495 controls, all heterozygous for DRB1*1501-DQB1*0602. We found association with a protective variant near HLA-DQA2 (rs2858884; P < 3 x 10(-8)). Further analysis revealed that rs2858884 is strongly linked to DRB1*03-DQB1*02 (P < 4 x 10(-43)) and DRB1*1301-DQB1*0603 (P < 3 x 10(-7)). Cases almost never carried a trans DRB1*1301-DQB1*0603 haplotype (odds ratio = 0.02; P < 6 x 10(-14)). This unexpected protective HLA haplotype suggests a virtually causal involvement of the HLA region in narcolepsy susceptibility.