DNA palette code for time-series archival data storage

Abstract

The long-term preservation of large volumes of infrequently accessed cold data poses challenges to the storage community. Deoxyribonucleic acid (DNA) is considered a promising solution due to its inherent physical stability and significant storage density. The information density and decoding sequence coverage are two important metrics that influence the efficiency of DNA data storage. In this study, we propose a novel coding scheme called the DNA palette code, which is suitable for cold data, especially time-series archival datasets. These datasets are not frequently accessed, but require reliable long-term storage for retrospective research. The DNA palette code employs unordered combinations of index-free oligonucleotides to represent binary information. It can achieve high net information density encoding and lossless decoding with low sequencing coverage. When sequencing reads are corrupted, it can still effectively recover partial information, preventing the complete failure of file retrieval. The in vitro testing of clinical brain magnetic resonance imaging (MRI) data storage, as well as simulation validations using large-scale public MRI datasets (10 GB), planetary science datasets and meteorological datasets, demonstrates the advantages of our coding scheme, including high net information density, low decoding sequence coverage and wide applicability.