Tran, Nhu QuynhKoçak, MehmetMendeş, Mehmet2025-01-272025-01-2720171308-78942146-8877https://doi.org/10.5336/biostatic.2016-53627https://search.trdizin.gov.tr/tr/yayin/detay/293386https://hdl.handle.net/20.500.12428/17100Objective: The objective of this study is to compare performances of maSigPro, BATS,Repeated Measures ANOVA (RM-ANOVA), and Classical RM-ANOVA, which are commonlyused methods, for testing interaction effect in time-course microarray experiments. Material andMethods: We generated random time-course gene expression profiles from multivariate normaldistribution based on two different covariance structures, three different time-series lengths, andtwo different profile types. Results: The results suggested that the BATS and the maSigPro aregenerally more superior methods to the ANOVA-based methods in terms of sensitivity and specificitymeasures. In general, the BATS was more appropriate for medium and especially for longterm data sets while maSigPro was appropriate almost for all experimental conditions. On theother hand, the ANOVA-based methods had a good performance especially for the short-termtime-course profiles. Conclusion: Thus, we concluded that length of time-course measurementand profile types are two major factors affecting the performances of these methods.eninfo:eu-repo/semantics/openAccessBiyolojiİstatistik ve OlasılıkArticle91354410.5336/biostatic.2016-53627293386