Application of principal component analysis for gene sequences

In this study, principal component analysis has been applied on data comprising of 6675 gene and 20 sequence collected by using cDNA microarray technology from livers of mice used in toxicology studies in certain time periods. cDNA microarray analysis is an efficient technology used for simultaneously analyzing of thousands of genes obtained from multiple experiments or samples. As for principal components analysis, it is a multivariable statistical technique used for the purpose of dimensional minimizing in order to attain the minimal number of principal components representing the whole data structure and explaining variance-covariance structure of original variables and eliminating the dependency structure. Forming of gene groups from similar expression profiles and description of related genes which are implemented by similar component loads among the groups have been explained by using this cDNA technology. Besides that, interpretation and decomposition of factors (components) from correlation matrix which belongs to same data group have been explained. Some of the methods developed for minimizing the data set to fewer components which can explain the whole data structure have been evaluated. According to Scree Graph, one of these niinimizing methods, components until the region where the curve starts to become straight are accepted. In other words, principal components beyond the point where the curve starts to become a straight line can be declined (ignored), therefore, it can be concluded that 9 or 10 eigen values would be enough according to this method. Besides that, Methods suggested by Kaiser and Barlett have also been taken into consideration. If we assume that the first 10 eigen values are enough to describe the whole variance, then in this case, it is thought that it is good enough to describe the whole variance by using 10 eigen values with a variance loss of 17.372% instead of describing the whole variance by using 20 eigen values. Key Words: cDNA microarrays, Gene expression, Genetic analysis, Principal Component Analysis n