Fig. 4: Dynamic, adaptive sampling leads to improved SNP discovery.
From: Dynamic, adaptive sampling during nanopore sequencing using Bayesian experimental design
We compared the variants called from data collected by the control (dashed lines) and by BOSS-RUNS (solid lines) to ground truth variants from deep, short-read sequencing of the same strains. Performing variant discovery at different timepoints gives further insight into the advantages of our method. a, Whereas the sensitivity of BOSS-RUNS is slightly lower for the most abundant species, we observe a larger number of discovered true positives in all remaining genomes. To highlight differences, we set the y-axis ranges to 0–0.95 for the first three species and 0–0.035 for the remaining two. b, The precision of variants called from data generated using BOSS-RUNS is at a similar level to the control or moderately higher.
