Beyond the Alphabet: Deep Signal Embedding for Enhanced DNA Clustering

2024-10-08

arXiv.org

AI Summary

To address the bottleneck of inaccurate read assignment in DNA storage caused by multiple repetitive sequences, this paper proposes a novel end-to-end clustering paradigm operating directly on raw nanopore electrical signals—bypassing conventional basecalling, which introduces error accumulation and information loss. Methodologically, we design a deep embedding network integrating Transformer architectures and contrastive learning, coupled with a differentiable clustering module to jointly optimize signal-level similarity metrics. Experiments on real nanopore sequencing data demonstrate a 12.7% improvement in clustering accuracy and a 3.2× speedup in inference time over traditional base-sequence-based approaches. To our knowledge, this is the first work to shift clustering from the basecall domain to the raw signal domain, establishing a new pathway toward high-accuracy, high-throughput DNA data decoding.

Technology Category

• Machine Learning: Clustering
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Search and Optimization: Learning to Search

Application Category

• Security and Privacy: Data transparency and provenance
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Web Mining and Content Analysis: Normalization, clustering, classification, and summarization of Web text

Abstract

The emerging field of DNA storage employs strands of DNA bases (A/T/C/G) as a storage medium for digital information to enable massive density and durability. The DNA storage pipeline includes: (1) encoding the raw data into sequences of DNA bases; (2) synthesizing the sequences as DNA extit{strands} that are stored over time as an unordered set; (3) sequencing the DNA strands to generate DNA extit{reads}; and (4) deducing the original data. The DNA synthesis and sequencing stages each generate several independent error-prone duplicates of each strand which are then utilized in the final stage to reconstruct the best estimate for the original strand. Specifically, the reads are first extit{clustered} into groups likely originating from the same strand (based on their similarity to each other), and then each group approximates the strand that led to the reads of that group. This work improves the DNA clustering stage by embedding it as part of the DNA sequencing. Traditional DNA storage solutions begin after the DNA sequencing process generates discrete DNA reads (A/T/C/G), yet we identify that there is untapped potential in using the raw signals generated by the Nanopore DNA sequencing machine before they are discretized into bases, a process known as extit{basecalling}, which is done using a deep neural network. We propose a deep neural network that clusters these signals directly, demonstrating superior accuracy, and reduced computation times compared to current approaches that cluster after basecalling.

Problem

• DNA storage
• grouping accuracy
• efficient retrieval

Innovation

• Improved DNA Storage
• Nanopore Sequencing
• Deep Neural Networks