Skip to main content
. 2020 Dec 7;37(9):1211–1217. doi: 10.1093/bioinformatics/btaa953

Fig. 2.

Fig. 2.

(a) Overview of preparation of training datasets using ONT’s retraining model, Taiyaki. Labeled reads obtained by Taiyaki are then split into fixed-length raw signals and corresponding nucleotide sequences. (b) Overview of evaluation of different basecallers in terms of SNV-detection performance assuming short-read sequencing as the ground truth