SAN DIEGO (GenomeWeb) – RNA sequencing can help improve the diagnostic rate of exome and genome sequencing-based tests for genetic disease, according to a new study presented at the American Society for Human Genetics' annual meeting here this week.
Despite the success of exome and genome sequencing to uncover the molecular cause of genetic disorders, laboratories have reported diagnostic rates ranging from 25 to 50 percent, leaving the majority of cases unsolved, David Murdock, assistant director of the clinical lab at Baylor College of Medicine's Human Genome Sequencing Center, said in a presentation.
As part of the National Institutes of Health's Undiagnosed Diseases Network, Murdock said, he and others wanted to see whether RNA sequencing could help solve some of the cases that were initially negative.
As a first step, the group analyzed gene expression levels of known disease genes in controls to determine whether looking for expression of these genes in blood and fibroblasts would be sufficient.
Next, they analyzed 66 cases via a targeted RNA-seq panel that focused on genes associated with 165 previously identified variants whose effect was unknown. Around 60 percent of those variants were intronic or deep intronic and were thought to impact splicing; 24 percent were null variants; 10 percent were missense variants; and 6 percent were other types of variants, including synonymous ones.
RNA sequencing strongly contributed to the diagnosis in 14 cases, or 21 percent, Murdock said. It was also important for ruling out variants as pathogenic. Specifically, RNA-seq was able to rule out 42 percent of the previously identified variants, including 90 percent of the intronic and deep intronic variants, for which it indicated that they had no effect on splicing.
Notably, RNA-seq found that 60 percent of the synonymous variants did affect splicing, which is important because "it's difficult to ascribe these to pathogenicity," he said.
RNA sequencing solved the most cases — a total of eight — in individuals who had a neurological disorder, demonstrating that despite not analyzing the affected tissue, RNA-seq could still be informative.
Murdock described a specific example, a case of a three-year-old girl with developmental delay, dysmorphic features, and cardiac defects. Her phenotype resembled Kabuki syndrome, but clinical trio exome sequencing had been negative prior to enrolling in the Undiagnosed Diseases Network. After she enrolled, the researchers first analyzed the previous sequencing data and identified a de novo deep intronic variant. Although in silico algorithms predicted that the variant could have an impact on splicing, there was not hard evidence to call it as anything other than a variant of unknown significance. The variant was also in a gene for which other de novo variants have been shown to be associated with a Kabuki-like syndrome. RNA sequencing found that the variant created a new splicing site that was not present in controls. Murdock said that although the overall expression level did not change, one theory is that the "abnormal protein may not function correctly." Additional work needs to confirm that, he said, but "we think this contributes to the diagnosis."
Murdock added that while the targeted RNA-seq approach helped improve the diagnostic rate, there were still a number of challenges. Most notably, gene expression is often tissue specific, so although blood can act as a good proxy, some genes are not expressed at high enough levels to detect them. In addition, "better tools are needed for a transcriptome-wide approach," he said, since this study evaluated just variants that had previously been identified.
Nonetheless, he said, the study "supports the inclusion of RNA-seq in genetic workups." This would be particularly important as more diagnostic sequencing protocols are transitioning from exome to whole-genome sequencing, where more noncoding variants with unknown functions will be identified. He added that it would also be important for groups like the American College of Medical Genetics and Genomics to develop guidelines for interpreting variants from RNA-seq data.
In another presentation at ASHG, Samya Chakravorty said that his team at Emory University tested the ability of RNA sequencing to aid with the diagnosis of patients suspected to have a muscular dystrophy known as dysferlinopathy. The disorder is often undiagnosed because it has phenotypes that are similar to other myopathies. A precise molecular diagnosis is important for this disorder, he said, because there are biomarker-based clinical trials.
Chakravorty's group first used a blood monocyte assay to test for the dysferlin protein in 350 individuals suspected of having the disorder, followed by RNA sequencing and a 35-gene panel.
RNA sequencing helped improve the diagnostic rate from 37 percent to 80 percent for the 350 patients. In addition, RNA sequencing helped clarify cases for which variants of unknown significance were identified.
In particular, Chakravorty said, RNA sequencing helps to understand important gene domains that harbor variants that are used to stratify patient into clinical trials. In addition, there is evidence that some cases of dysferlinopathy may be impacted by variants in genes other than the known disease gene DYSF. For instance, Chakravorty's team found a variant in the gene MYH2 in one individual with rapid disease progression. "We're doing more functional studies to tease this apart," he said.