Figure 2

Statistical power analysis. Statistical power (probability, y-axis) to detect significant strand bias at a human exome position with Fisher's exact test, depending on read coverage (x-axis) and on assumed position-specific error rate (color) higher than the assumend background error rate of 0.01. Example: Even at an extremely high error rate of 0.5 (cyan), even a coverage of 100 grants only a discovery chance of 40%. Fluctuations are caused by finite sampling size for simulations.