Random access DNA memory in a scalable, archival file storage system

Abstract

DNA is an ultra-high-density storage medium that could meet exponentially growing 1 worldwide data storage demand. However, accessing arbitrary data subsets within exabyte-2 scale DNA data pools is limited by the finite addressing space for individual DNA-based 3 blocks of data. Here, we form files by encapsulating data-encoding DNA within silica 4 capsules that are surface-labeled with multiple unique barcodes. Barcoding is performed 5 with single-stranded DNA representing file metadata that enables Boolean logic selection on 6 the entire pool of data. We demonstrate encapsulation and Boolean selection of sub-pools of 7 image files using fluorescence-activated sorting, with selection sensitivity of 1 in 10 6 files per 8 channel. Our strategy in principle enables retrieval of targeted data subsets from exabyte-9 and larger-scale data pools, thereby offering a random access file system for massive 10 molecular data sets. 11 12 DNA is the polymer used for storage and transmission of genetic information in biology. In 13 principle, DNA can also be used as a medium for the storage of arbitrary digital information at 14 densities far exceeding existing commercial data storage technologies and at scales well beyond 15 the capacity of current data centers 1 . Ongoing advances in nucleic acid synthesis and sequencing 16 technologies also continue to reduce dramatically the cost of writing and reading DNA, thereby 17rendering DNA-based digital information storage potentially viable economically in the near 18 future 2-5 . As demonstrations of its viability as a general information storage medium, to date 19 books, images, computer programs, audio clips, works of art, and Shakespeare's sonnets have all 20 been stored in DNA using a variety of encoding schemes 6-12 . In each case, digital information was 21 converted to DNA sequences and typically fragmented into 100-200 nucleotide (nt) blocks of data 22

Autonomous Computing Materials

“… Because these DNA-based quantum dot and dye networks evolve in their theoretical underpinnings and their capabilities are implemented experimentally, key questions we seek to explore include how to extract complex behavior and learning from large-scale in vivo neuronal recordings; how to encode behavior and learning within these networks; how to program robustness and error correction, which are hallmarks of the brain; how to integrate multimodal sensing, recording, and computational abilities; and how to facilitate these capabilities in an autonomous manner that consumes minimal-to-no energy, while also operating in nonconventional wet, physiological, and other uncontrolled environments. In parallel, synergistic work, we are using DNA as an information-coding polymer itself, with a density that exceeds 10 18 bits per cubic millimeter, random-access capabilities, − and a shelf life that can be extended to millennia using silica encapsulation …”

Data Storage Based on DNA

“…Though PCR‐based random access is very convenient to amplify specific DNA sequences, accessing arbitrary data subsets within exabyte scale DNA data pools by this method is limited because of the finite addressing space for individual DNA sequence. For this reason, many new strategies based on DNA hybridization, [ 76 ] physical isolation, [ 59 ] molecular interactions, [ 77,78 ] and fluorescence‐activated sorting [ 79 ] for random access have been developed. Concretely, Plesa et al [ 76 ] synthesized oligos using microarray‐based synthesis and exposed a single‐stranded region that acts as a gene‐specific microbead barcode for pulling down the complementary oligos to assemble each gene ( Figure a).…”

“…Tomek et al [ 78 ] used the primers functionalized with chemical handles in PCR to create chemically labeled copies of different desired files’ strands and selected the desired file's strands with functionalized magnetic beads (Figure 6c). Recently, Banal et al [ 79 ] encapsulated DNA datasets within silica capsules that were addressed with multiple unique barcodes. To query a file or several files, fluorescently labeled 15‐mer ssDNA probes that are complementary to file barcodes were added to the data pool.…”

Encoding scheme for data storage and retrieval on DNA computers

“…The volume of information that can be combined today is constrained for the most part by the expense of combination and sequencing, however, development in the biotechnology industry forecasts requests of extent cost decreases and proficiency upgrades. Banal et al [21] recently presented a strategy that could provide random access to large archival files. Since they encapsulated the data in silica particles, the density of data storage is higher than PCR‐based methods.…”