Fig. 4: Enlarged epi-bit data storage and data retrieval by one-pot sequencing.
From: Parallel molecular data storage by printing epigenetic bits on DNA
a, Preprocessing of the original image data. b, Preparation of barcoded DNA carriers. c, Writing of epi-bit information onto several DNA carriers. d, Stored images of DNA bases and derivatives. 5hmC, 5-hydroxymethylcytosine; 6mA, N6-methyldeoxyadenosine; 7mG, N7-methylguanine. e, Images recovered by called methylation probabilities from nanopore sequencing reads (greyscale). f, Binarized image recovery based on the threshold of methylation calling. Coloured bars below each image indicate the indices of carriers used to encode the image. g, Determination of the threshold for methylation calling. The threshold was chosen to minimize the total overlap (potential miscalling) between dual-normal distributions fitted to methylation calling data at each epi-bit site. h, k-means clustering of all sequencing reads based on called epi-bit patterns. Patterns on each cluster are mapped to the original coded templates. i, Total barcode mixing fractions (representing read misalignment errors) as a function of the hamming distance between pairs of epi-bit patterns. The simulated (sim.) data (red circles) fitted to y = exp(−0.5x + 3.8). The experimental (exp.) data (blue circles) fitted to y = exp(−0.48x + 3.7) + 0.9.
