High-Density and Low-Coverage Data Storage in DNA Using Efficient Coding Strategies

We design efficient error correction coding and constrained coding strategies that are specifically tailored to the dominant errors and biochemical characteristics of phosphoramidite chemistry synthesis and Illumina sequencing. We also perform large-scale in-vitro DNA data storage experiments to validate our proposed coding architecture. In particular, random datasets of 1.54 MB and 1.61 MB were each encoded into 30,000 oligos, achieving net information densities of 1.731 and 1.815 bits/nt, respectively. All stored information was successfully recovered without error in 1,000 random downsampling experiments at average sequencing coverages of only 4.5x and 6.0x, respectively.