TABLE 2

Effect of sample size on misclassifications

Sample sizeRibosomal amplificationAvg % FP ± SDaAvg % FN ± SDa
5,000 (1×)bLow18.4 ± 0.620.2 ± 0.7
5,000 (1×)Medium17.3 ± 0.719.4 ± 0.7
5,000 (1×)High17.8 ± 0.720.4 ± 0.7
50,000 (10×)Low14.7 ± 0.616.2 ± 0.6
50,000 (10×)Medium13.2 ± 0.614.5 ± 0.6
50,000 (10×)High15.3 ± 0.616.9 ± 0.6
500,000 (100×)Low14.1 ± 0.614.5 ± 0.7
500,000 (100×)Medium9.2 ± 0.6)9.8 ± 0.5
500,000 (100×)High14.0 ± 0.614.7 ± 0.6
5,000,000 (1,000×)Low10.8 ± 0.710.8 ± 0.6
5,000,000 (1,000×)Medium8.3 ± 0.68.4 ± 0.6
5,000,000 (1,000×)High14.1 ± 0.714.1 ± 0.7
50,000,000 (10,000×)Low10.7 ± 0.610.5 ± 0.7
50,000,000 (10,000×)Medium8.1 ± 0.68.1 ± 0.6
50,000,000 (10,000×)High14.0 ± 0.614.0 ± 0.7
  • a The error rates are expressed as percentages of the populations that were detected in both DNA and RNA sampling profiles, and each value is the average of 100 independent iterations ± the standard deviation). With low sampling depths, the number of detected populations was less than the 5,000 input populations.

  • b The parenthetical factors indicate the sampling depth as a multiple of the number of input populations (5,000) in the simulated community.