Composite Hedges Nanopores: A High INDEL-Correcting Codec System for Rapid and Portable DNA Data Readout

Abstract

DNA, as the origin for the genetic information flow, has also been a compelling alternative to non-volatile information storage medium. Reading digital information from this highly dense but lightweighted medium nowadays relied on conventional next-generation sequencing (NGS), which involves wash and read cycles for synchronization and the indel (insertion and deletion) errors rarely occur. However, these time-consuming cycles hinder the future of real-time data retrieval. Nanopore sequencing holds the promise to overcome the efficiency problem, but high indel error rates lead to the requirement of large amount of high-quality data for accurate readout using emerging NGS-based codec systems. Here we introduce Composite Hedges Nanopores (CHN), a nanopore-based codec scheme tailored for real-time data retrieval, capable of handling indel rates up to 15.9% and substitution rates up to 7.8%. The overall information density can be doubled from 0.59 to 1.17 by utilizing a degenerated eight-letter alphabet, where one composite strand will be projected into eight normal strands. We demonstrate that sequencing times of 20 and 120 minutes were sufficient for processing representative text and image files (7 and 115 composite strands), respectively. The time-diminishing deviations are mainly originated from the extremely uneven abundance among the composite strands (cross-group variation) as well as the huge inequality among the normal strands (in-group variation). Moreover, to achieve complete data recovery, it is estimated that text and image data require 4x and 8x physical redundancy (coverage) of composite strands, respectively. Our CHN codec system excels on both molecular design and equalized dictionary usage, laying a solid foundation for nucleic acid-based data retrieval and encoding approaching to real-time, applicable in both cloud and edge computing systems.