Abstract
BackgroundChoosing the appropriate sample size is an important step in the design of a microarray experiment, and recently methods have been proposed that estimate sample sizes for control of the False Discovery Rate (FDR). Many of these methods require knowledge of the distribution of effect sizes among the differentially expressed genes. If this distribution can be determined then accurate sample size requirements can be calculated.ResultsWe present a mixture model approach to estimating the distribution of effect sizes in data from two-sample comparative studies. Specifically, we present a novel, closed form, algorithm for estimating the noncentrality parameters in the test statistic distributions of differentially expressed genes. We then show how our model can be used to estimate sample sizes that control the FDR together with other statistical measures like average power or the false nondiscovery rate. Method performance is evaluated through a comparison with existing methods for sample size estimation, and is found to be very good.ConclusionA novel method for estimating the appropriate sample size for a two-sample comparative microarray study is presented. The method is shown to perform very well when compared to existing methods.
Highlights
Choosing the appropriate sample size is an important step in the design of a microarray experiment, and recently methods have been proposed that estimate sample sizes for control of the False Discovery Rate (FDR)
In this article we introduce a mixture model approach to estimating the underlying distribution of noncentrality parameters, and effect sizes, for t-statistics observed in pilot data
For reasons discussed above we focused on controlling average power, along with the positive false discovery rate (pFDR), in our sample size estimates
Summary
Choosing the appropriate sample size is an important step in the design of a microarray experiment, and recently methods have been proposed that estimate sample sizes for control of the False Discovery Rate (FDR). Many of these methods require knowledge of the distribution of effect sizes among the differentially expressed genes. If this distribution can be determined accurate sample size requirements can be calculated. In the case of replicated two-sample comparisons statistical tests may be used to assess the significance of the measured differential expression. A successful approach to this has been to control the false discovery rate (FDR) [2], or FDR-variations like the positive false discovery rate (pFDR) [3]
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.