Fig. 6

Sources of variance in the total RNA mixture dataset. An analysis of variance in R [17] was used to identify the major sources of variability in measuring 4 RNA mixture pairs by two laboratories (see Fig. 2, pathways 5 and 6). A four-way ANOVA, testing all possible interactions, was applied to determine the contribution to variance by analyte, mixture, isolation process, and PCR laboratory. In panel a, the input data for the model was the Cq for each of the miRNA measured by each PCR laboratory for each mixture prepared from four isolations (see Fig. 5, panel a). In panel b, the input data was the ΔCq between each mixture from the same dataset. The mean of squares is plotted on the y-axis on a log10 scale and is a measure of the average contribution of each factor or interaction to the variability in the experiment. Factors or interactions with a mean of squares less the one are not included. White bars correspond to biological factors or interactions, black bars correspond to laboratory factors or interactions, and grey bars correspond to interactions between biological and laboratory factors