ISSN:
1573-9686
Source:
Springer Online Journal Archives 1860-2000
Topics:
Medicine
,
Technology
Notes:
Abstract A statistic is defined from the analysis of variance of replicate measurements in biological time series data. This statistic, the square root of the mean square of the variation within groups of two replications, is termed the index of replication,S Δ, and is a measure of the precision of an assay.S Δ was found to be a better indicator of assay replicability than the conventional index of assay precision, λ, calculated from duplicate standard curves. When a power spectrum permutation test, as defined elsewhere, is used to detect a significant periodicity in discrete, amplitude-noisy, time series data, the relationship betweenS Δ and the 95% confidence limit on peak-to-peak signal variation shows that the variation must be greater than 1.4S Δ for detection. Since the value of 1.4S Δ assumes duplicate determinations per point and a uniform distribution of measurement error, it is a conservative estimate of the strength of the power spectrum permutation test to detect one frequency. When a conventional analysis of variance is used, the level of 1.4S Δ is not reached until quadruplicate determinations have been processed; thus, the present approach is more efficient.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1007/BF02363559
Permalink