Israeli scientists have achieved a major advancement in DNA data storage technology, dramatically accelerating the speed at which information can be retrieved from DNA strands.
A research team at Technion - Israel Institute of Technology developed an artificial intelligence system called DNAformer that can read DNA-encoded data 3,200 times faster than existing methods, while also improving accuracy by up to 40 percent.
The AI model, based on a transformer architecture and trained on synthetic data, can reconstruct accurate DNA sequences even from error-prone copies. It incorporates a specialized error-correction algorithm optimized for DNA-based storage.
In testing, the researchers encoded a 3.1-megabyte dataset containing an image, audio clip, text about DNA storage, and random data. Using DNAformer, they retrieved this information in just 10 minutes - a process that previously took several days. The system achieved a data density of 1.6 bits per DNA base even under challenging noise conditions.
While DNA storage shows immense potential for handling the world's growing data needs - projected to reach 180 zettabytes by 2025 - slow retrieval speeds have been a major obstacle to practical implementation. This breakthrough brings the technology closer to real-world viability.
The Technion team plans to continue refining DNAformer for different storage applications. The system is designed to be scalable and adaptable, with researchers already considering how to optimize it for future advances in DNA sequencing technology and market requirements.
Though still slower than conventional digital storage methods, this dramatic speed improvement represents a major step forward in making DNA-based data storage practical for future use. The technology cleverly combines nature's time-tested data storage mechanism with cutting-edge artificial intelligence to help address humanity's expanding data storage needs.