Each disorder had previously been mapped to a chromosomal locus and candidate gene sequencing failed to identify the pathogenic variant. For six disorders, an average autozygous block of 4.4 Mb contained only one novel homozygous variant and rendered disease gene identification straightforward. Even in the case of infantile parkinsonism-dystonia syndrome, for which autozygosity and linkage mapping were only partially informative, a managable list of 19 candidate variants was assembled simply by assuming mutation homogeneity. Prior mapping data and thorough knowledge of the patient implicated a single variant from this list. For four of the conditions, more than one affected individual was available for exome sequencing. Even in the absence of mapping data, the identification of the putative pathogenic variant would still have been unambiguous. When we examined the shared, novel homozygous variants in affected individuals, we found one, and only one, that was not homozygous in any other individuals in the study. Thus, the assumption of mutation homogeneity obviates the need for SNP genotyping and mapping; we reach the same conclusion by exome sequencing of multiple affected individuals without the added time and expense of SNP genotyping. The number of novel homozygous variants in each individual was surprisingly small. Average inbreeding coefficients of 4% and 2.5% in the Lancaster Amish and Mennonite populations, respectively, suggested that a small but significant fraction of variation will be homozygous. On average, we found only 21 novel homozygous variants per sample across the exome. Of these variants, only 12 were predicted to be potentially pathogenic. This represents 3.7% of all novel variants per exome. For the two AMN107 disorders where a singleton was sequenced, we identified only 6 potentially pathogenic novel variants which were homozygous in the patient but in no other samples. Since our total sample size was small, we expect that future studies which leverage accumulated exome data will allow us to sequence single individuals to identify rare, uniquely homozygous pathogenic variants. In the outbred population, a R428 strategy that scans for homozygosity or compound heterozygosity for novel variants in the same gene should yield equally manageable candidate gene lists. Among the fifteen individuals studied, we found 4200 different novel autosomal sequence variants, roughly 62% of which have pathogenic potential. We infer that 3.6% of these variants were non-pathogenic changes as they were homozygous in one or more unaffected individuals. As more Amish and Mennonite exomes are analyzed within a clear clinical context, our ability to determine pathogenicity will improve. The exome data also provided a broader view of the genetic disease burden within these populations.