Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore a groundbreaking conference presentation that introduces SUSTag-ORCtrL, an innovative DNA tag system designed to enable real-time selective sequencing and address the critical challenge of random access in DNA storage applications. Learn how this novel approach combines de novo designed DNA tags (SUSTag) with an optional reject deep learning model and transfer learning (ORCtrL) to minimize crosstalk and achieve efficient random access capabilities. Discover the technical details of SUSTag's design methodology, which utilizes Bhattacharyya distance methods and incremental clustering to create up to 384 DNA tags with minimal interference, achieving an impressive intra-class F1 score of 99.05% that surpasses existing tagging systems. Understand how ORCtrL integrates optional-reject CNN-LSTM architecture with transfer learning to significantly improve accuracy from Q21 to Q25 on training data and enhance pre-trained model performance from Q9 to Q13 on new datasets. Examine the practical implementation using Composite Hedges Nanopores encoding for storing text and image data with SUSTag as address bits, demonstrating PCR-free random access with nearly 100% access purity and data decoding capabilities within three hours. Gain insights into the potential applications of this technology for genomics and transcriptomics while advancing practical, high-throughput DNA information retrieval systems for long-term archival storage solutions.
