In 2018, Frances Arnold highlighted the limitations of current DNA technology in composing whole genes or genomes, attributing this ability solely to evolution.

Patrick Hsu, Brian Hie, and their team of scientists aimed to overcome this by training an AI model to mimic evolutionary processes.

The team utilized 300 billion DNA letters from 80,000 genomes to recreate cellular complexity.

Their research delves into the potential of generative AI for biological data analysis and genome manipulation.

Megan Molteni, a science writer at STAT, reported on these groundbreaking advancements in genomic medicine and reproductive technology.