Korea Develops 'World's Best' Next-Generation Gene Editing Design Technology

Yonsei University Research Team, "Efficient Design of Prime Editor Possible with AI"
Released Worldwide, Available for Anyone to Use Freely

Korean researchers have developed a technology that can efficiently design next-generation gene scissors that eliminate the possibility of side effects such as genetic mutations using artificial intelligence (AI).

The Ministry of Science and ICT announced on the 29th that Professor Hyungbeom Kim's research team at Yonsei University has developed an AI model that can precisely and safely design the next-generation gene scissors called prime editors. The research results attracted international attention, being published on the same day in the international academic journal 'Cell'.

An illustration of gene editing technology using enzymes to cut specific gene regions.

This technology is gaining attention as a next-generation gene scissors technology that can complement the shortcomings of CRISPR-Cas9, one of the widely known gene editing technologies.

Gene scissors are tools that use enzymes to easily cut desired gene regions. Recently, next-generation gene scissors called prime editors have been developed, which can not only cut genes but also correct them by including new genetic information to fill the cut sites. Prime editors are noted as a much safer technology than existing gene scissors because they cut only one strand of the DNA double helix instead of completely cutting both strands.

The problem is that prime editors are structurally more complex and have more diverse cases than previous gene scissors because they include new genetic information. Therefore, it is difficult to design precise and safe gene scissors. For clinical research and therapeutic purposes, it is necessary to accurately understand under what conditions optimal gene editing occurs without errors, and for this, securing and analyzing large-scale experimental data is essential.

The research team secured more than 330,000 prime editor data points over about three years since 2020 and experimentally measured the efficiency of each prime editor. This is the largest scale of measurement data reported in academia so far. Through this, they succeeded in revealing the main mechanisms determining the performance of prime editors and their effects. By undergoing a learning process through artificial intelligence with the analyzed data, they created a model that predicts the efficiency of usable prime editors when inputting gene sequence information to be corrected.

The model developed this time shows world-class prediction performance regarding the precision and safety of prime editors, and researchers worldwide can use this model via the internet. Previously, designing prime editors required experimenting with dozens to hundreds of prime editors each time, but using the technology implemented by the research team this time, prime editors can be designed quickly and accurately, which is expected to have high utility in the future gene therapy field.

Professor Hyungbeom Kim stated, “This is the most outstanding predictive model reflecting all known elements of prime editors so far,” and added, “Now gene editing using prime editors has become easy to utilize.”