Skip to main content
Fig. 2 | International Journal of Bipolar Disorders

Fig. 2

From: Detecting significant genotype–phenotype association rules in bipolar disorder: market research meets complex genetics

Fig. 2

Illustration of the implemented version of the association rule mining algorithm. The lattice shown left is traversed starting from root {} to all leaves. Each genotype pattern (node in the tree) represents a subgroup of patients shown in the genotype matrix G. Additionally, using the p phenotype information of the patients from matrix P, we can count genotype and phenotype occurrences in contingency tables. Here illustrated for the genotype pattern g1g2gn with ‘a’ counting all patients where genotype g1g2gn and phenotype pi are present, ‘d’ were neither of both are present, and ‘b’ and ‘c’ counting patients with presentation of genotype g1g2gn but not phenotype pi and visa versa. The lattice is traversed as long as there are unprocessed genotype patterns that cannot be pruned before

Back to article page