Multilocus Association Analysis Into Susceptibility Genes Of Schizophrenia

Screening the susceptibility genes of complex diseases which based on mathematics, computer science, information technology and molecular biology is the frontier of bioinformatics. It adopts modern mathematics, computer and information technology to research, makes use of large amounts of data generated by the research of molecular biology, genomics and proteomics, to search,discovery and identify complex disease-related genes. The SNP not only is the markers of the individual differences and the human race, but also it decides the markers of the susceptibility of diseases and sensitivity of medications, and it provides a basis of screening the susceptibility genes and diagnosing the diseases and selecting medicine.Schizophrenia is a serious mental disorder with a lifetime prevalence rate of 1% in the general population worldwide. For this reason, confirming the pathogeny, preventing and curing schizophrenia on the gene level, is a hotspot of the medical research. There are two steps for mapping a disease-related gene in the human genome, the linkage-based genome-wide scan and the regional mapping with the association analysis. In this paper, the detailed method was to define the 200 trios composed of a schizophrenia patient and their healthy parents from Han nationality in China as research targets, using Case-Parental Control method, discussing the relationship between schizophrenia and the polymorphism of 13 SNP of cPLA2 family genes and 5 SNP of PNPLA8 gene. This research will break a new method for the genetic theory of Schizophrenia.The kernel of this paper is the statistic analysis, basing on the genotype data from the former medicine experiments. By using R language, for the first time the multilocus association analysis based on the merging algorithm is used to research schizophrenia. As well as the study estimates the missing data, examines the Hardy-Weinberg equilibrium, linkage disequilibrium and haplotype significance test. The methods have the following features: Firstly, the multilocus association analysis is used, which is more efficient than the traditional method based on the single-marker. Now when analyzing these data of SNP the methods are limited to traditional single SNP method, e.g. hapoltype relative risk test (HRR) and transmission disequilibrium test (TDT). When taking use of these methods, the statistical significance is poor and it is hard to get an explicit conclusion sometimes. Secondly if using the traditional medicine experiments completely, it must be waste of time and energy, the study overcomes the shortcomings and uses a new computer model by bioinformatics, the merging algorithm based on variable length Markov chains. The computer model automatically builds a tree-structured model of genomic multilocus data, and then produces a parsimonious model by the merging algorithm, at the same time tests for association between trait data and graph-edge counts, in order to judge whether there is any relationship between gene Polymorphism and disease. Thirdly, differing from the traditional case-control study, the case-parental control design method avoids population stratification when parents used as controls, splits the parental haplotypes into 200 haplotypes transmitted to the affected offspring (the"case"haplotypes) and 200 untransmitted haplotypes (the"controls"). Fourthly, the study makes use of the program PHASE based on a Bayesian statistical method to estimate the missing data. When collecting the samples, part of genotype data might be missing for some reasons. If discarding all these missing data, a large quantity of information will be destroyed. The estimation insures the veracity and integrity.The process is as follows. I used PHASE to estimate missing data and unresolved phase. The Hardy-Weinberg equilibrium was tested. The alleles of the parents in the trios were split into 200 cases and 200 controls. I used Language R to do multilocus association analysis based on the merging algorithm, and obtained the positive loci primarily. I further analyzed the positive loci, combined these loci with others on the same chromosome into haplotypes, then I test the association for haplotype significance and linkage disequilibrium.The conclusions are as follows. Besides PLA2G4C(rs2303744) and PNPLA8(rs6466240), the goodness of fit test showed that genotype frequency distributions of the other 16 loci were not deviated from the H-W equilibrium, thus these samples were suitable for the genetic analysis. In total samples, the single-marker association tests show that there was no significant difference for the frequencies of alleles of these 16 SNP between the case and the control(P>0.05).This indicated no association between the 16 SNP and schizophrenia. The estimated LD showed that rs2396001 and rs10249427,rs2396001 and rs40848,rs2396001 and rs40876,rs10249427 and rs40848 were in the same LD block with SHEsis programs. Analysis for haplotype significance test showed that four haplotypes had association with schizophrenia. rs3816533(T)-rs4924595(C)-rs1668589(C)-rs1356410(G),rs891014(T)-rs156631(C)-rs2307279(C),rs2162886(T)-rs891014(T)- rs156631(C)- rs2307279(C),rs2396001(C)-rs10249427(T)-rs40848(C)-rs40876(G).In a word, these findings above suggest thoroughly that the multilocus association analysis based on the merging algorithm is more efficient than the single-marker association test.The paper is on the basis of the molecular experiments of schizophrenia susceptibility genes, applying the effective multilocus association analysis method. Not only this method can guide the medical test , but may reduce the invalid experiments and time. This study was very important for elucidating the etiology and genetic mechanisms of schizophrenia at a molecular level, and also for the development of genetic diagnosis, new drugs for the treatment of the illness and prediction of schizophrenia risk. The current study would also provide study strategy and precious experience for other complex genetic disorders.Unifying multi-disciplinary and developing the bioinformatics is a new tendency in biology research. The new methods and the new viewpoints of schizophrenia from the scholars in different field must certainly lead the research of complex diseases into a new time.