We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.IT

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Information Theory

Title: Concatenated Code Design for Constrained DNA Data Storage with Asymmetric Errors

Abstract: DNA Data storage has recently attracted much attention due to its durable preservation and extremely high information density (bits per gram) properties. In this work, we propose a hybrid coding strategy comprising of generalized constrained codes to tackle homopolymer (run-length) limit and a protograph based low-density parity-check (LDPC) code to correct asymmetric nucleotide level (i.e., A/T/C/G) substitution errors that may occur in the process of DNA sequencing. Two sequencing techniques namely, Nanopore sequencer and Illumina sequencer with their equivalent channel models and capacities are analyzed. A coding architecture is proposed to potentially eliminate the catastrophic errors caused by the error-propagation in the constrained decoding while enabling high coding potential. We also show the log likelihood ratio (LLR) calculation method for the belief propagation decoding with this coding architecture. The simulation results and the theoretical analysis show that the proposed coding scheme exhibits good bit-error rate (BER) performance and high coding potential ($\sim1.98$ bits per nucleotide).
Subjects: Information Theory (cs.IT); Other Quantitative Biology (q-bio.OT)
Cite as: arXiv:2112.09417 [cs.IT]
  (or arXiv:2112.09417v1 [cs.IT] for this version)

Submission history

From: Yixin Wang [view email]
[v1] Fri, 17 Dec 2021 10:25:00 GMT (784kb,D)

Link back to: arXiv, form interface, contact.