Gene ranking through the integration of synchronization experiments

Identification of genes expressed in a cell-cycle-specific periodical manner is of importance and has attracted significant interest recently. However, the identification of cell cycle regulated genes by microarray gene analysis is complicated by both synchronization loss and the presence of noise, and remains an interesting challenge. It is known that periodicity and regulation are two vital components observed in popularly used synchronization experiments-alpha arrest, cdc15, and cdc28. We hypothesize that an integrated analysis of aligned synchronization experiments can improve the identification of cell-cyclic genes of the Saccharomyces cerevisiae (budding yeast) data set. In this paper, we propose a unique ranking scheme based on an integrated analysis of two synchronization experiments (cdc15 and cdc28) for a gene using Pearson correlation coefficients and principle component analysis (PCA). Skewness and kurtosis based features vectors are also discovered for the correlated information space. Genetic algorithm based classification employed for the identification of known cell cycle genes from the discovered features of the ranked genes results in higher degrees of accuracy.