Channel models and error correction codes for DNA information storage

摘要

It has been demonstrated in recent years that synthetic DNA can be used to reliably store large volumes of data. It should be possible to recover the data from the synthetic DNA after very long time periods under fairly mild storage conditions. Two key requirements are the need to avoid repeated symbols known as homopolymers and the need to avoid errors arising from secondary structures. In this paper, an error model is developed and error correction techniques are proposed for this technology. The use of variable length Huffman codes in the avoidance of homopolymers can lead to loss of synchronisation if any errors do occur. A scheme to recover synchronisation is proposed and shown to be effective.