HEDGES error-correcting code for DNA storage corrects indels and allows sequence constraints

Abstract

Synthetic DNA is rapidly emerging as a durable, high-density information storage platform. A major challenge for DNA-based information encoding strategies is the high rate of errors that arise during DNA synthesis and sequencing. Here, we describe the HEDGES (Hash Encoded, Decoded by Greedy Exhaustive Search) error-correcting code that repairs all three basic types of DNA errors: insertions, deletions, and substitutions. HEDGES also converts unresolved or compound errors into substitutions, restoring synchronization for correction via a standard Reed–Solomon outer code that is interleaved across strands. Moreover, HEDGES can incorporate a broad class of user-defined sequence constraints, such as avoiding excess repeats, or too high or too low windowed guanine–cytosine (GC) content. We test our code both via in silico simulations and with synthesized DNA. From its measured performance, we develop a statistical model applicable to much larger datasets. Predicted performance indicates the possibility of error-free recovery of petabyte- and exabyte-scale data from DNA degraded with as much as 10% errors. As the cost of DNA synthesis and sequencing continues to drop, we anticipate that HEDGES will find applications in large-scale error-free information encoding.

Robust data storage in DNA by de Bruijn graph-based decoding

“…al. implemented a hash encoded, decoded by greedy exhaustive search (HEDGES) codes that can correct indels with a single copy of DNA strand 11 . Despite outstanding technical achievements in indel correction with single copy data, HEDGES requires 40% redundancy codes for correction of merely 3% indel errors.…”

Decoding DNA data storage for investment

“…Despite inherent error rates, DNA data storage can in principle tolerate high error rates in both write and read channels through sufficient redundancy, appropriate codecs (coder-decoder), error correction codes and algorithm design ( Erlich and Zielinski, 2017 ; Organick et al, 2018 ; Press et al, 2020 ).…”

“…Beside the substitution errors that are found in the latter, nucleotide insertions or deletions are additional common error types that occur during DNA synthesis and sequencing. One such algorithm has recently been developed to repair all three error types, where insertions or deletions can be corrected directly within a single DNA strand, unlike previous codes that correct substitution errors ( Press et al, 2020 ). A part of this algorithm, known as HEDGES (Hash Encoded, Decoded by Greedy Exhaustive Search), translates between a string of the four nucleobases and a binary string of the digital binary code (see 3.1.…”

Robust retrieval of data stored in DNA by de Bruijn graph-basedde novostrand assembly

“…al. implemented a hash encoded, decoded by greedy exhaustive search (HEDGES) codes which can correct indels with a single copy of DNA recently 11 . Despite HEDEGES shows outstanding technical achievements, the coding rate of HEDGES drops below 0.6 for correction of 3% errors.…”

“…Increasing indels will lead to an overwhelming number of "substitutions" exhausting the column EC codes 8,10 . Recently developed hash encoded, decoded by greedy exhaustive search (HEDGES) codes can be used to correct higher frequency of indels 14 . Based on the simulation, HEDGES codes require 40% redundancy codes for correction of 3% indel errors, which would significantly increase the cost of data writing.…”