Recent genome scans have established the presence of a major psoriasis-susc
eptibility locus in the human leukocyte antigen (HLA) complex on chromosome
6p21.3. To narrow the interval for candidate gene testing, we performed a
linkage-disequilibrium analysis of 339 families, with the use of 62 physica
lly mapped microsatellite markers spanning the major histocompatibility com
plex (MHC). As detected by use of the transmission/disequilibrium test (TDT
), individual markers yielded significant linkage disequilibrium across mos
t of the MHC. However, the strongest evidence for marker-trait disequilibri
um was found in an similar to 300-kb region extending from the MICA gene to
the corneodesmosin gene. Maximum-likelihood haplotypes were constructed ac
ross the entire MHC in the original sample and across a 1.2-Mb region of th
e central MHC in an expanded sample containing 139 additional families. Sho
rt (two- to five-marker) haplotypes were subjected to the TDT using a "movi
ng-window" strategy that reduced the variability of TDT P values relative t
o the single-locus results. Furthermore, the expanded sample yielded a shar
p peak of evidence for linkage disequilibrium that spanned similar to 170 k
b and that was centered 100 kb telomeric to HLA-C. The 1.2-Mb interval was
further dissected by means of recombinant ancestral haplotype analysis. Thi
s analysis identified risk haplotype 1 (RH1), which is a GO-kb fragment of
ancestral haplotype 57.1, on all identifiable HLA risk haplotypes. One of t
hese haplotypes exhibits significant linkage disequilibrium with psoriasis
but does not carry Cw6, which is the HLA allele most strongly associated wi
th the disease. These results demonstrate that RH1 is highly likely to carr
y the disease allele at PSORS1, and they exclude HLA-C and corneodesmosin w
ith a high degree of confidence.