| 摘要 |
The present invention provides methods for analyzing measurement errors in measured signals obtained in an experiment, e.g., measured intensity signals obtained in a microarray gene expression experiment. In particular, the invention provides a method for transforming measured signals into a domain in which the measurement errors in the transformed signals are normalized by errors as determined from an error model. The methods of the invention are particularly useful for analyzing measurement errors in signals in which at least portion of the error is dependent on the magnitudes of the signals. Such transformed signals permit analysis of data using traditional statistical methods, e.g., ANOVA and regression analysis. Magnitude-independent errors can also be used for comparing level of measurement errors in signals of different magnitudes.
|
