Variability as a function of sequence abundance.

<p>The forward direction is indicated in black, the reverse direction in grey. A) Variability in the 100 most common sequences from a non-homogeneous clinical sample run multiple times (N = 47 replicates over 18 deep sequencing runs). B) Variability between the forward and reverse direction for samples in MOTIVATE/A4001029 dataset (N = 1521 samples; 41030 unique sequences, median 26 [IQR 19–34] per sample). Less abundant sequences tended to have more variable prevalence estimates in comparison to more abundant sequences.