Table 1.

Estimation of the number of detected transcripts at 30 million records.

Dataset	Predicted FPKM > 0	Percent of whole
Bladder	97 872-105 433	83%-84%
Hypoxia	100 613-102 279	90%-93%
nisc2	101 490-110 594	71%-76%
Oncopig	26 865-30 658	87%-92%

Abbreviation: FPKM, Fragments Per Kilobase of transcript per Million mapped reads.

For each run of Figure 1, a linear model of the number of detected transcripts was created using the formula detected ~log(records) (R² = 0.988-1.0). These models were then used to predict the number of detected transcripts at 30 million records for each run. As each dataset contains between 4 and 7 runs, this table lists the range of predicted transcripts for each dataset. The values for percent of whole were calculated by dividing the predicted number of transcripts at 30 million records by the number of detected transcripts measured in the full dataset as plotted in Figure 1.