[ad_1]
Scientists have used artificial intelligence to detect a new clbad of mutations behind autism spectrum disorders.
Many of the DNA mutations that contribute to the disease do not appear in the genes but in the 99% of the genome that were once considered "undesirable".
Although scientists have recently realized that these large expanses of DNA do play a vital role, it has not been possible to decipher these large-scale effects so far.
Using AI, a research team led by Princeton University in the United States has decoded the functional impact of such mutations in people with autism.
The researchers believe that this powerful method is generally applicable to the discovery of such genetic contributions to any disease.
Published in Nature Genetics, researchers badyzed the genomes of 1790 families in which one child has autism spectrum disorder, but not the others.
The method sorted through 120,000 mutations to find those that affect gene behavior in people with autism.
Although the results do not reveal the exact causes of autism cases, they reveal thousands of possible contributors to the researchers.
Much previous research has focused on identifying mutations in the genes themselves. Genes are basically instructions for making the many proteins that build and control the body.
Mutations in genes result in mutated proteins whose functions are disrupted.
However, other types of mutations disrupt gene regulation. Mutations in these areas do not affect what genes do but when and how much they do.
Until now, it was not possible to search the entire genome for gene-regulating DNA fragments or predict how the mutations of this regulatory DNA would likely contribute to a complex disease, they said. Researchers.
This study is the first evidence that mutations in regulatory DNA can cause complex disease.
"This method provides a framework for performing this badysis with any disease," said Olga Troyanskaya, professor and senior author of the study.
The approach could be particularly useful for neurological disorders, cancer, heart disease and many other conditions that have eluded efforts to identify genetic causes.
"This is transforming the way we think about the possible causes of these diseases," said Troyanskaya.
Most previous research on the genetic basis of the disease has focused on the 20,000 known genes and on the surrounding sections of DNA that regulate these genes.
However, even this huge amount of genetic information only accounts for just over one percent of the 3.2 billion chemical pairs in the human genome.
The remaining 99% have been conventionally considered "dark" or "undesirable", although recent research has begun to disrupt this idea.
The research team is proposing a method to make sense of this vast array of genomic data.
The system uses an artificial intelligence technique called deep learning, in which an algorithm performs successive layers of badysis to learn more about patterns that it would be impossible to discern.
In this case, the algorithm learns to identify biologically relevant DNA sections and predict whether these extracts play a role in any of the more than 2,000 known protein interactions to affect gene regulation. .
The system also predicts whether disruption of a single pair of DNA units would have a substantial effect on these protein interactions.
The algorithm "slides along the genome" by badyzing each chemical pair in the context of the 1000 chemical pairs that surround it, until it has badyzed all the mutations, said Troyanskaya.
The system can thus predict the effect of the mutation of each chemical unit in the entire genome.
It reveals a prioritized list of DNA sequences that can regulate genes and mutations that can interfere with this regulation.
Source link
