Comparison of Genehunter 1.2 max LOD, max Heterogeneity LOD, and max NPL statistics for disease gene localization using simulated data. C. Chen¹, D. Gordon², S. Finch³, N. Mendell³. 1) Case Western Reserve University, Cleveland OH; 2) Rockefeller University, New York NY; 3) State University of New York at Stony Brook, Stony Brook NY.

   This study compares the performance of the maximum LOD(MLOD), maximum heterogeneity LOD(MHLOD) and maximum GENEHUNTER 1.2 Nonparametric Linkage(MNPL) score for disease-gene localization. We use a simulation study considering: (i)4 disease models, (ii)100 replicates for each model, (iii) 58 pedigrees whose structure matches a field study with 592 subjects (iv)3 marker loci with 3 equally frequent alleles each, (v)marker distances of 10cM between each marker locus (vi)disease locus placed halfway between the second and third marker and (vii)either 0% unlinked families (linkage homogeneity) or 50% unlinked families (linkage heterogeneity). The disease models have: (i) a disease allele, A, with frequency 0.03, penetrance of the disease allele for those homozygous for A of 0.8, penetrance of the disease allele for those homozygous for the normal allele, B, of 0.01 and (ii)either complete dominance or complete recessivity of the disease allele. For each replicate we obtain the marker with MNPL, the marker with MLOD, and the marker with MHLOD and compute the distance of these markers from the disease locus. The MLOD (and MHLOD) are obtained by maximizing LOD (or HLOD) over penetrances, and allele frequencies, and recording the maximum of these two scores.
   Comparison of distances between the location with the MLOD, MHLOD and MNPL values based on our 15-replicate pilot study indicates that (i)for data simulated under linkage homogeneity, the same location is identified by both the MLOD and MHLOD statistics; under linkage heterogeneity the location of MHLOD is much closer to the true position of the disease locus than the position of MLOD (ii) for either linkage homogeneity or heterogeneity, the position indicated by MHLOD is closer on average to the disease locus than the location identified by MNPL. Findings for the remaining 85 replicates will be presented at the meeting.