Fig. 2: De novo construction of non-canonical DeepSME basecaller between Alice and Bob. | Nature Communications

Fig. 2: De novo construction of non-canonical DeepSME basecaller between Alice and Bob.

From: De novo non-canonical nanopore basecalling enables private communication using heavily-modified DNA data at single-molecule level

Fig. 2

a Training DeepSME with a three-stage alignment-free pipeline. b 9-mer coverages of three used training datasets. c Training error rate as a function of standard deviation of segmented lengths. Inset depicts the uncertainty (standard deviation of a certain segmented length). d Mean absolute current difference (MACD) of both canonical cytosine (top) and 5hmC (bottom) at different nucleotide positions of a k-mer oligonucleotide. e Methylated 5hmC vs. canonical cytosine of normalized current levels. f Recall vs. precision of DeepSME for both canonical and methylated 5hmC. a contains icons from Bootstrap Icons (https://icons.getbootstrap.com/), licensed under CC BY 3.0 (https://creativecommons.org/licenses/by/3.0/).

Back to article page