Data Analysis

To test diversity estimates for accuracy, for each of the four validation distributions (a-d), we calculated the deviation as $\delta = (\hat{S} - S)/S$, where $S$ is true diversity and $\hat{S}$ estimated diversity. We calculated mean and standard deviation of $\delta$ from 20 replicate samples in which 500 individuals were sampled before calculating estimators, to emulate the process of sampling random clones in preparing a cDNA library. To identify reliable estimators, two two-tailed univariate tests were performed of the null hypothesis that $\delta = 0$, namely Student's t, and the non-parametric Wilcoxon signed-rank test, with $\nu =19$.

P-values were interpreted to indicate significant differences at 95% and 99% experiment-wide confidence levels. The correction to maintain experiment-wide confidence for twenty multiple comparisons was $\alpha'=\alpha/20$, where $\alpha$ is the size of the test.

Methods in R, version 1.1.1 [62] performed the tests (t.test and w.test).

We used the most accurate, least-biased estimators, as judged by this experiment, to infer diversity in empirical transcript libraries.