A.I Finds Autism Causing Mutations

This is the first known study to functionally link mutations in regulatory DNA with complex diseases and suggests that the alterations affect expression of genes in the brain including those responsible for neuron migration and development. According to the researchers this approach could be used more widely to study roles of noncoding mutations in disorders such as heart disease and cancer.

Progress has been made in understanding the genetics of autism spectrum disorder, such studies suggest mutations in protein coding genes only account for about 30% of spontaneous cases of ASD where there is no known family history. Only 1-2% of the human genome comprises genes that code for protein the rest of the genome is noncoding DNA historically known as junk DNA, that we now know contains regulatory regions playing important roles in controlling where/when genes are expressed.

Roles for noncoding mutations have been speculated in ASD, but it has not been possible to go through the entire genome to identify alterations in regulatory DNA to predict how these changes may contribute. Any one individual change may have dozens of noncoding mutations, which will all be unique, and identifying common mutations among those with the same disorder has not been possible.

“Analysis of contribution of noncoding mutations to ASD is challenging due to the difficulty of assessing which noncoding mutations are functional and, of these, which contribute to the disease phenotype.”

Taking a different approach than past studies a machine deep learning model was trained how to predict how a variation in a stretch of noncoding DNA might affect gene expression, this model was then applied to Simons Simplex Collection autism population comprised of whole genomes of close to 2,000 family quartets including a child with autism, and unaffected sibling, unaffected parents, and no prior family history of ASD; meaning any mutations in the affected child occurred spontaneously and were likely to have played a role in the disorder.

The algorithm was applied to data from 1,790 quartets, where it learned patterns in the genome and taught itself how to identify biologically relevant sections of DNA, and how to predict whether alterations in noncoding regions may play a role in any of the 2,000+ protein interactions affecting gene regulation to predict the effect of any mutation in the genome and generate a prioritized list of DNA sequences/disease impact score likely to regulate genes and mutations likely to impact that regulation.

“What our paper really allows you to do is take all those possibilities and rank them. That prioritization itself is very useful, because now you can also go ahead and do the experiments in just the highest priority cases.” says Christopher Park, PhD.

“This is a shift in thinking about genetic studies that we’re introducing with this analysis. In addition to scientists studying shared genetic mutations across large groups of individuals, here we’re applying a set of smart, sophisticated tools that tell us what any specific mutation is going to do, even those that are rare or never observed before.”  says Chandra Theesfeld, PhD.

Mutations in the noncoding regions are suggested to affect similar genes and function to those previously linked to ASD in studies on coding genes. The authors write: “Notably, our study reveals important biological convergences among the genetic dysregulations associated with ASD. Our analyses of the disease impact of mutations with effects on DNA and RNA point to similar sets of impacted genes and pathways, indicating that the effects of regulatory mutations are convergent. Furthermore, high-impact noncoding regions that we find in ASD probands affect the same genes previously found to be impacted by LoF [loss of function] coding mutations in ASD … This convergence provides support for a causal contribution of noncoding regulatory mutations to ASD etiology.”

“This is consistent with how autism most likely manifests in the brain. It’s not just the number of mutations occurring, but what kind of mutations are occurring.” according to Park.

“Right now, 98% of the genome is usually being thrown away. Our work allows you to think about what we can do with the 98%.” adds Troyanskaya.