| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 5845073 | Progress in Neuro-Psychopharmacology and Biological Psychiatry | 2012 | 8 Pages |
We attempted to estimate how many genes are involved in schizophrenia using a simulation based on the polygenic threshold model. The basic assumptions were as follows: (1) All genes involved are transmitted independently; (2) every locus is composed of two alleles - one pathogenic and the other non-pathogenic; (3) all pathogenic alleles are dominant; (4) the two alleles at any locus are in Hardy-Weinberg Equilibrium (HWE) in the general population (GP) but not within the patient (PP) or non-patient (NP) subpopulations; (5) the number of affected loci determines the disease genetically; and (6) only a fraction of genetically determined individuals actually becomes ill. A range of the total number of disease-related genes (N) and threshold genetic load (T) was set for the simulation. Assuming that the number of affected loci follows a binomial distribution, the mean gene frequencies satisfying a disease prevalence of 1.12% in the GP were sought for various N and T combinations. Based on these gene frequencies, the odds ratio and the incidence rate in relatives under random mating were calculated. These results were then compared with real genetic epidemiologic data to obtain best-fit estimates for N and T. The results indicated that a polygenic threshold model with an N greater than 100 and a T in the range of 0.3-0.8 fits the empirical data. It was estimated that at least several hundreds of study subjects are required to yield a statistically significant frequency difference for a single gene between the patient and the control groups.
⺠Genetics of schizophrenia was simulated using the polygenic threshold model. ⺠Odds ratio and prevalence in relatives were calculated by the simulation. ⺠The model with more than 100 genes and genetic heterogeneity fits real world data.
