Skip to main content
Fig. 2 | Genome Biology

Fig. 2

From: Comprehensive benchmark and architectural analysis of deep learning models for nanopore sequencing basecalling

Fig. 2

Latest basecallers benchmark on the human task. a Basecalling reads failure rates: pass, failed mapping, short alignment, and no prediction (from top top bottom). b Alignment event rates: match, mismatch, insertion, and deletion (from top to bottom). c Homopolymer error rates per base. d PhredQ scores distributions for correctly (light) and incorrectly (dark) predicted bases. e AUC of match rate sorted by average read PhredQ score

Back to article page