An Improved Marker Code Scheme Based on Nucleotide Bases for DNA Data Storage

Abstract

Due to the rapid growth in the global volume of data, deoxyribonucleic acid (DNA) data storage has emerged. Error correction in DNA data storage is a key part of this storage technology. In this paper, an improved marker code scheme is proposed to correct insertion, deletion, and substitution errors in deoxyribonucleic acid (DNA) data storage. To correct synchronization (i.e., insertion and deletion) errors, a novel base-symbol-based synchronization algorithm is proposed and used. In the improved scheme, the marker bits are encoded as the information part of the LDPC code, and then mapped into marker bases to correct the synchronization errors. Thus marker bits not only assist in regaining synchronization, but also play a role in LDPC decoding to improve decoding performance. An improved low-complexity normalized min-sum (INMS) algorithm is proposed to correct residual substitution errors after regaining synchronization. The simulation results demonstrate that the improved scheme provides a substantial performance improvement over the concatenated marker code scheme and concatenated watermark code scheme. At the same time, the complexity of the INMS algorithm was reduced, while its bit error rate (BER) performance was approximate to that of the belief propagation (BP) algorithm.

Keywords

• DNA data storage
• insertion/deletion
• low-density parity-check (LDPC) codes
• marker codes
• synchronization