Skip to main content
Premium Trial:

Request an Annual Quote

RNA-Seq Data Helps Focus Search for Rare Variants

NEW YORK – RNA sequencing data can be used to help find rare genetic variants, according to a new study. The method, which looks at imbalances in expression from the two alleles of a gene, could be helpful in diagnosing rare diseases, the authors said.

The idea is to use RNA expression of different alleles as a functional readout of the genome, said Pejman Mohammadi of the Scripps Translational Science Institute, the study's first author. When considering rare diseases, rare genetic variants are usually present on one copy. "We looked at a bunch of individuals for each gene to see how much variation between the parental copies is ok. We have that for all specific genes." If, say, the maternal copy is more expressed compared to the paternal copy, "we could assume either one has a mutation that changed expression," he said.

The process doesn't necessarily say which variant may be responsible for a disease. But it could highlight genes for further analysis. In the paper, published today in Science, the researcher showed how they were able to apply the method to RNA-seq data from 70 Mendelian muscle disease patients, resolving previous cases and leading to one confirmed new diagnosis.

"This provides a new type of tool to home in on what are the potential causal genes" of some diseases, said Stephen Montgomery, a professor of pathology at Stanford University's Bio-X. Montgomery was not involved in this study, but is collaborating with Mohammadi on a different paper. "It's important because we know there are a number of rare genetic variants but it's very hard to identify which ones are functional."

Allele-specific expression, also called allele imbalance, has been known about for some time and it has previously been used to identify rare variants and their effects. But the new paper is the first "to really come up with a method to be able to identify those types of patterns more rigorously," Montgomery said.

Most genes have equal expression between the alleles, but some don't and that's perfectly normal, Mohammadi said. But by applying to allele-specific expression analysis to many individuals from the Genotype-Tissues Expression data set, the researchers were able to determine the genetic variation that could be expected. Any outliers from that, rather than using a threshold, might return results with rare variants.

"It's an interesting idea to use the GTEx data to help identify mendelian disease genes," said Joe Gleeson, a professor at the University of California, San Diego and Howard Hughes Medical Institute Investigator, who researches the genetics of pediatric brain diseases. "It's a massive data set and people are interested in understanding its relevance to human health."

But the method may be limited to the data its working from. "This concept was demonstrated to be useful with muscle disease, which happens to be a tissue sampled by GTEx, but whether this same concept would apply to tissues not sampled from GTEx is unknown and how well it would apply to other classes of disease would also be unknown," Gleeson said.

Montgomery added that the method only works with a subset of genes that are well-expressed — approximately half of all genes expressed in a particular cell

Mohammadi noted that the method still requires follow-up. "If you want to actually find the final variants, you'd need another downstream analysis," he said. Even in the study, the researchers didn't follow up with all the undiagnosed patients. "It takes a while to prove what [the disease-causing variant] actually is," he said.

But Mohammadi said his team is now working with hospitals, including Rady Children's Hospital in San Diego, to further refine the method. Montgomery noted, "we've used this for rare disease samples that span a wide number of different phenotypes, everything from neurological to muscle disorders."

"Really it's all comers. It can have a broad impact in terms of different types of rare disease," Montgomery said.