Integrating Genomic Data Sets for Knowledge Discovery: An Informed Approach to Management of Captive Endangered Species

Figure 4

Predictive modeling of disease prevalence and allele frequencies. The value in using genomic information to predict individuals at risk for an undesirable clinical phenotype depends upon the prevalence of the undesirable phenotype (disease) in the total captive population, the prevalence of the susceptibility genotype (exposed), and the prevalence of the disease in the individuals with the susceptibility genotypes (disease in exposed group). The matrix in the figure provides example values for a hypothetical endangered captive population of 100 individuals for which a disease phenotype has a population prevalence of 30%. The abbreviations “E+” and “” correspond to exposed and nonexposed, respectively. In this model, the exposed group has the allele/genotype(s) associated with the undesirable phenotype, while the nonexposed group does not have the susceptibility allele. Genotype-phenotype predictions, based on drastic SNP occurrence in genes associated with comparative genomics phenotype annotations derived from knockout mouse models, allow classification of members of the endangered species population into either a susceptible or less-susceptible class. Through bidirectional communication among zoo veterinarians, SSPs, zoo staff, and genomics scientists, genotype-phenotype predictions may be validated. Threshold values for increased relative risk in the exposed group, along with threshold levels of allele/genotype frequencies in the E+ and E− groups, will affect the success in employing such an approach for captive species management, as will the mode of inheritance (e.g., autosomal recessive versus autosomal dominant).