Particle.news
Download on the App Store

Rice Scientists Publish CLASSIC to Massively Expand DNA Circuit Design and Testing

A peer-reviewed Nature paper details a workflow linking full DNA circuit sequences to cell behavior at unprecedented scale by combining long- with short-read data plus machine learning.

Overview

  • The method generates hundreds of thousands to millions of genetic circuit designs in a single build, far exceeding prior scales.
  • It pairs long-read sequencing for complete circuit maps with high-accuracy short-read barcodes to connect each design to fluorescence readouts in engineered human embryonic kidney cells.
  • The resulting datasets train machine-learning models that predict performance of untested circuits, with 40 of 40 subsequent lab checks matching the predictions.
  • Researchers observed many distinct designs achieving the same function, suggesting flexibility that could support more robust engineered cells.
  • Authors say the scale enables the first accurate AI-guided analysis of gene circuits, pointing to faster routes to cell-based therapies, though the work remains at a research stage.