Scientists from Shanghai have made a breakthrough in genetic engineering by creating a new tool for DNA editing. Using artificial intelligence, researchers modified a tiny protein called Fanzor, derived from forest fungi, transforming it into a highly precise genome editor named enFanzor. The study results were published in the prestigious journal Nature Communications.
From Nature to the Lab: The Optimization Challenge
The natural Fanzor protein initially attracted the attention of geneticists due to its compact size — it is half the size of the standard CRISPR system, which consists of more than 1300 amino acids. However, in its natural form, this protein functioned poorly in human cells, rendering it useless for medical practice.
Instead of the traditional and lengthy manual screening of mutations, scientists entrusted the optimization process to a computer model called Fanzor-Fitness Predictor. The neural network analyzed patterns in related natural proteins and, without resorting to preliminary laboratory tests, selected the best options for point substitutions.
Results of the Algorithms
A combination of five top mutations proposed by the AI produced a colossal effect: the efficiency of editing human genes increased from 3% to 36% — almost 12-fold. However, the work did not stop at the protein alone. The second stage involved reducing the size of the system's "navigator" — the guide RNA, which directs the protein scissors to the specific DNA section.
The original guide RNA consisted of 350 nucleotides, making it too bulky for viral transport and expensive to produce. After filtering by the neural network and reinforcing the structure, the length of the molecule was reduced to 75 "letters." This reduction amounted to almost 80%.
Unexpected Compression Effect
Compressing the guide produced an unexpected side effect: enFanzor began working significantly more precisely, virtually eliminating "off-target" erroneous cuts in the genome. Furthermore, the compact guide RNA allowed for the creation of an efficient base editor based on Fanzor for the first time. This tool is capable of performing delicate rewriting of a single DNA letter without breaking the chain itself; meanwhile, the accuracy of point correction increased from 7% to 33%.
The new editor proved its competitiveness, competing on equal terms with standard CRISPR, and outperformed other existing miniature analogues in 22 out of 24 cases.
Tests on Living Organisms
The true power of the technology was demonstrated in tests on living biological models. In experiments with human stem cells, the tool successfully edited nearly 50% of blood cells, targeting the genetic switch responsible for hereditary diseases such as severe thalassemia. For comparison: the wild, unmodified version of the Fanzor protein in the same test barely reached 1% efficiency.
In experiments on mouse embryos, scientists introduced enFanzor to block the pigmentation gene. The result exceeded expectations — efficiency exceeded 90%, and offspring consisting of 12 completely white mice were born, each receiving the genetic change.
The Future of Gene Therapy
Researchers are convinced that this method opens new horizons. Now, computer models can modernize hundreds of other inefficient natural proteins in a matter of days, turning them into compact medicines against severe genetic diseases that will be easy to deliver into patients' bodies.
