Statistical models application and algorithm development for genetic studies in neuropsychiatric diseases

Author(s)

Rodrigo Secolin

• Document type: Doctoral Thesis
• Press: Campinas, SP.
• Institution: Universidade Estadual de Campinas (UNICAMP). Faculdade de Ciências Médicas
• Defense date: 2011-08-02

Examining board members

Íscia Teresinha Lopes Cendes; Edi Lúcia Sartorato; Mônica Barbosa de Melo; Aguinaldo Luiz Simões; Ana Lúcia Brunialti Godard

• Advisor: Íscia Teresinha Lopes Cendes

Abstract

Genetic factors have been described for several central nervous system diseases. A main step for disease gene identification is genetic mapping study. In addition, due new genotype acquire technology, the development of genotype processing data software is required. The objectives of this work were: 1) to generate interface between genotype equipment and statistical software by processing data algorithm; 2) to apply and evaluate statistical models in family sample segregating three neurological diseases: mesial temporal lobe epilepsy (MTLE), bilateral perysilvian polymicrogyria (BPP) and bipolar affective disorder (BPAD). Data interface was developed from a logic algorithm, which adds a genotype matrix data from equipment to a family data matrix. This algorithm, called JINGLEFIX, was implemented in PERL computer language and R environment. In addition, this software was used in genetic mapping study for MTLE, BPP and BPAD. Segregation analysis was performed in 148 nuclear MTLE pedigrees, with a total of 698 individuals, since this syndrome has not known inheritance pattern. One BPP pedigree with known X-linked dominant pattern of inheritance, with a total of 15 individuals, was submitted to parametric linkage analysis by LINKAGE package, using 18 microsatellite markers on candidate region Xq27-Xq28. Non-parametric linkage analysis was performed from 74 BPAD families, with a total of 411 individuals, by transmission disequilibrium test (TDT) and using 21 single nucleotide polymorphisms (SNPs) for 21 candidate regions. Segregation analysis revealed a major effect gene with an autosomal dominant pattern of inheritance and minor gene effect, which could influence MTLE phenotype. Further whole genome analysis mapped the putative MTLE major gene on 18p11. Parametric linkage analysis mapped Xq27 locus for BPP, a different region compared to the Xq28 previous described. This difference could be explained to sample type used by the two studies. Non-parametric linkage for BPAD identified the candidate region on 3p22. Further studies using 94 additional SNPs on 3p21-3p22 and gene expression analysis identified ITGA9 as susceptibility gene for BPAD. A comparison of statistical power between statistical analyses showed a high statistical power for parametric linkage analysis from one or a few large families; whereas a high statistical power was observed for non-parametric linkage analysis using several moderate size families. The conclusion of this study is that data processing algorithm and adequate statistical model applying are fundamental tools for successful of genetic mapping of complex diseases (AU)