Mendelian Diseases, Dominant
Mendelian diseases with dominant modes of inheritance pose greater technical challenges for genetic analysis. Heterozygous variants are generally more difficult to detect and analyze, not only due to their sheer number as compared to homozygous variants but also because they are subjected to higher false positive and false negative errors. Nevertheless, one could still design one of the following experiments to elucidate the genetic architecture underlying these diseases, depending on the availability of patient cohorts and the effect of the disease on reproductive fitness.

Familial cohort
If large families with the disease of interest are available, WES should be performed on at least two affected family members. In analyzing the results, any shared variants could potentially be disease-causing; using linkage data to focus on the variants located in linked intervals would greatly reduce the number of variants to be considered. A recent example of such an approach is a WES study on familial amyotrophic lateral sclerosis (FALS) [6]. The authors recruited two large families with FALS and selected two affected members with the greatest genetic distance from each family for WES. Among the rare functional variants that were identified, they tested the shared variants for Mendelian segregation among affected family members, eventually narrowing their search down to a single common gene, PFN1.

Non-familial cohort
When large families with the disease are not readily available, one could perform WES on a number of unrelated patients with similar clinical manifestations and select genes that are commonly mutated in the patient cohort. In carrying out such an experiment, it is important to consider the background mutation burden of the entire human gene set; it is possible that the longest genes might contain the highest number of rare functional variants and therefore appear to be the most interesting. To circumvent such errors in analysis, the variant burden of each gene should be normalized using data from a control population. In a WES study of pseudohypoaldosteronism type II (PHAII), the variant burden of every gene from 11 unrelated patients was compared against that of 699 controls [7]. A single gene that was specifically enriched for mutations in the patient set, KLHL3, was identified (5 variants from 11 patients compared to 2 from 699 controls, with p-value of 1.1 × 10-8).

Non-inheritable diseases
If the disease phenotype is severe enough such that it affects reproductive fitness, one could hypothesize that de novo variants-variants that emerged during meiosis of the germ cells that are unique to the offspring-are directly associated with the disease. These can be detected by performing trio-based WES on the unaffected parents and the proband, and variants that are present in the offspring only and not in the parents will be called de novo variants. The rate of de novo mutations has been shown to be relatively stable and affected primarily by paternal age [8]. WES of family quartets can also be carried out, where unaffected siblings are recruited simultaneously and their de novo mutations are compared against those of the proband. A series of studies involving autism trios and quartets recently demonstrated that rare de novo mutations were associated with the risk of autism and identified multiple de novo mutations in SCN2A, KATNAL2, and CHD8, implicating these genes in the genetic etiology underlying autism [9-11].