Low-Bias Amplification for Robust DNA Data Readout

Abstract

11In DNA data storage, the huge sequence complexity is challenging repeatable and efficient 12 information reading. Here, we demonstrated that biased amplification is serious matter for stable 13 information retrieve from synthetic oligo pool comprising over ten thousand strands encoding over 14 2.85 MB digital information. Therefore, we adapted isothermal DNA amplification as a new 15 biochemical framework, termed as isothermal DNA reading (iDR) for low biased manipulation of 16 DNA pool with huge sequence complexity. iDR was built to work in a priming-free and error-17 spreading proof manner and achieved low-biased amplification. Finally, we demonstrated that 18 immobilized DNA materials read by iDR is able to stably read data deeply and repeatedly, the first 19 reported sustainable DNA storage system. Necessary amount of sequencing reads for perfect 20 2 oligos retrieve significantly decreased. These advanced features enable the iDR chemical 21 framework being ideal alternative for manipulating the huge sequence complexity and building 22 the sustainable and efficient DNA storage "hardware". 23 24 3 Introduction: 25 In DNA data storage, technology originally developed for bioengineering approach including array 26 oligo synthesis, PCR and DNA sequencing were integrated to construct the "hardware" for DNA 27 storage 1-5 . Array synthesized DNA oligo pool comprising of from thousands up to millions of oligo 28 strands with several hundred bases in length has been utilized in many applications, e.g., probe 29 blend, DNA origami assembly, and genome synthesis 1 . Because of both the location on microchip 30 and DNA sequence context, the copy number for each oligo strand from array synthesis is distinct 5 . 31Furthermore, the sheer number of synthesized oligos from array on microchip is very small, 32 roughly from 10 5 to 10 12 at the concentration of femtomolar depending on the synthesis platform 6, 33 7 . Generally, amount of DNA materials, over a few hundred nanograms, at the concentration of 34 micromolar is necessary for high quality DNA sequencing covering all oligos in the pool on 35 commercial DNA sequencing platform, e.g., Illumina 8 . Therefore, amplification is crucial to boost 36 the signal for the subsequent DNA sequencing for stably reading data. 37The oligo pool size for DNA data storage is much larger at least several orders of magnitude than 38 that in other bioengineering applications 9-11 . The unevenness of copy number generated a huge 39 complexity that caused serious problem for DNA molecule retrieval and data decoding 4, 5, 12, 13 . 40 Thus far, in reported systems the information reading was achieved by PCR amplification and 41 NGS, but the copy number unevenness originally stemmed from microchip synthesis was further 42 skewed by highly biased amplification process 5, 12 . Therefore, more amplified DNA materials was 43 necessary to fetch the minor oligo strands from the skewed oligo pool. The length and sequence 44 context, GC content and secondary structure of DNA molecule are well...

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

“…Previous studies have shown that the implementation of a reliable codec system for DNA data storage requires two layers of error correction (EC) codes. The outer layer codes handle the strand dropouts and the inner layer codes respond for the error-free decoding of DNA strands as many as possible [7][8][9][10] . Strand dropouts are erasures, which can be well solved by erasure codes, e.g.…”

Robust retrieval of data stored in DNA by de Bruijn graph-based de novo strand assembly

“…Frequently occurring errors in DNA synthesis, amplification, sequencing, and preservation, however, challenge the reliability of data storage in DNA. Efforts to solve this problem led to the implementation of a codec system which requires two layers of error correction (EC) codes, the outer layer codes handling the strand dropouts and the inner layer codes dealing with intramolecular errors, ensuring accurate information readouts 10- 13 . Strand dropouts are erasures and therefore they can be well solved by erasure codes, e.g.…”

“…None of the existing methods can be used to correct rearrangements. DNAs are subjected to hydrolysis and degradation 12,13 under certain conditions and in long-term storage, highlighting the importance of data stability studies on influences of harsh conditions, such as high temperature or UV exposure, and methods of DNA protection 8,14 . New coding methods that can tolerate fragmented DNA strands are nevertheless critical for enhancing the robustness of DNA data storage.…”