1. ^*Oxagen Limited, Abingdon, Oxon, UK
2. ^†Division of Genomic Medicine Mathematical Modelling and Genetic Epidemiology Group, Biomedical Genetics, University of Sheffield, Sheffield, UK
3. ^‡Division of Genomic Medicine, Biomedical Genetics, University of Sheffield, Sheffield, UK

Correspondence: Rachid Tazi-Ahnini, Division of Genomic Medicine, Biomedical Genetics, University of Sheffield, Medical School, Beech Hill Road, Sheffield S10 2RX, UK. Email: r.taziahnini@sheffield.ac.uk

¹Present address: Wales Gene Park, Medicentre Health Park, Cardiff CF14 4UJ, UK.

²Present address: Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden.

³Present address: Leicester & Warwick Medical School, University of Warwick, Coventry CV4 7AL, UK.

Received 25 May 2004; Revised 1 September 2004; Accepted 28 September 2004.

Abstract

Psoriasis is a chronic inflammatory skin disease that affects 0.1%–5% depending on the population. PSORS1 is the major susceptibility locus, accounting for approximately 33%–50% of the genetic component of psoriasis among Caucasians. PSORS1 is located within the major histocompatability complex (MHC) locus on 6p21.3. Its position has been refined to hundreds of kilobase and the region located at 100–200 kb telomeric to human leukocyte antigen (HLA)-C is a very strong candidate. To determine the MHC psoriasis risk haplotype, we screened the whole 46 kb interval for single-nucleotide polymorphisms (SNP) and identified 138 SNP. We genotyped 29 SNP throughout this region in psoriatic nuclear families. We calculated the frequency of haplotypes generated by the 29 SNP using all genotyped founder individuals and found four common haplotype with frequency >0.10. We then used SNPtagger to derive the best six SNP and fed these into Transmit using 148 nuclear families. We found that CTGGAC haplotype is a single-point score haplotypes telomeric to HLA-C and gives a 1 df, ² of 50.27 (p<0.0001). Most importantly the six selected SNP accurately tagged the most common haplotype found in this region. Moreover, using the same program (Transmit) we show that the association with CTGGAC is higher than the one with HLA-Cw6 (²=10.53; p=0.0051). Our results give scores as high as the highest single-point scores suggesting that it is unlikely to be able to discriminate the origin of the association on this analysis on strength of association.