Latent-Class Analysis ofRecurrence RisksforComplex Phenotypes withSelection andMeasurement Error: A TwinandFamilyHistory StudyofAutism

Summary Theuseofthefamily history method toexamine the pattern ofrecurrence risks forcomplex disorders such asautism isnotstraightforward. Problems suchasun- certain phenotypic definition, unreliable measurement with increased error rates formoredistant relatives, and selection duetoreduced fertility all complicate theesti- mation ofrisk ratios. Using datafromarecent family history study ofautism, andasimilar study oftwins, this paper showshowalatent-class approach canbe usedtotackle these problems. Newfindings arepre- sented supporting amultiple-locus model ofinheritance, withthree loci giving thebest fit.