More precision in DNA manipulation: AI helped with development
A new method should allow precise manipulation of large sections of DNA. As proof, the researchers produced herbicide-resistant rice.
(Image: Paul Sunderland/Shutterstock.com)
Researchers from the Institute of Genetics and Developmental Biology at the Chinese Academy of Sciences have reported a breakthrough in the precise manipulation of large sections of DNA. Their method is suitable for thousands to millions of base pairs and allows the editing of genetic material in plants and animals alike. Gao Ciaxia and his team published their research results online in the journal Cell.
The research is based on the Cre-Lox recombination system, which allows the targeted manipulation of DNA. Two sections of base pairs, the so-called LoxP sites, mark the DNA section to be modified. The recombinase known as Cre is an enzyme that docks onto the LoxP sites and removes or reverses the desired section, for example.
Manipulating DNA more successfully
According to the researchers, this promising method has several problems that the new procedures are intended to solve. One major issue with the method is the so-called reversible recombination reactions as undesirable effects. These ensure that the manipulations carried out on the genome are reversed and thus nullified. With their method, they were able to reduce the severity of this phenomenon by a factor of ten. They also developed a method to ensure seamless genome modifications. Conventional manipulation with the underlying method, on the other hand, is associated with a kind of scarring of the DNA.
Videos by heise
For their research, the team developed new recombinases using artificial intelligence. The enzymes created in this way showed a 3.5-fold higher recombination efficiency than the original wild type of the enzyme. The AI-supported method for constructing proteins (enzymes are usually proteins) was published by researchers from the same institute led by Gao Ciaxia just one month earlier in the journal Cell. AI models would generate compatible amino acid sequences based on three-dimensional protein structures, considering certain structural limits as constraints. Amino acids are the building blocks of proteins. This is how AI helps with efficient protein engineering.
To substantiate their research result in a concrete application, the researchers used it to produce herbicide-resistant rice. To accomplish this, they used their method to reverse engineer a 315 kilobase section of its DNA.
(dgi)
