The idea of doing studies on the association between HLA types and disease was first discussed around the mid-1960s, largely stimulated by Ruggero Ceppellini (a pioneer of early HLA studies who coined the word ‘haplotype’ in 1967) and based on the association between inherited blood disorders and malaria, a suggestion made by J.B.S. Haldane in 1949 (ref. 7). The first published study of an HLA and disease association was on Hodgkin's disease in 1967 (ref. 8). The claimed association was with an antigen then called ‘4c’. Even with the few antigens (about five) then ascertainable, the association (OR = 2.8, χ2 = 5.06) was not considered significant because of the problem of multiple comparisons. The study was, however, based on a good rationale—that is, on the association between Gross virus–induced leukemia and H-2 in the mouse9. The discovery of H-2–linked immune response genes by McDevitt and others soon provided the best explanation for such associations10. Although the association between Hodgkin's disease and HLA turned out to be relatively weak, with ORs < 1.5, it has been abundantly confirmed. Nevertheless, in
