Skip to main content
Premium Trial:

Request an Annual Quote

Jackson Lab Finds Structural Mutations Underlie Many Unsolved Exomes from Mendelian Disease Mice


NEW YORK (GenomeWeb) — Researchers at the Jackson Laboratory and their collaborators have found that exome sequencing identifies causative mutations in about half of Mendelian disease mouse models, and many of the unsolved cases appear to harbor structural mutations that whole-genome sequencing can detect.

The results, published online in Genome Research last month, suggest that whole-genome sequencing may be able to boost the diagnostic rate of clinical exome sequencing. In addition, findings in the mouse models may help with the interpretation of human clinical exome data.

The Jackson Laboratory breeds and provides more than 7,000 varieties of well-characterized mice to researchers across the world — last year, for example, it shipped about 2.5 million mice. In order to ensure that mouse strains remain genetically the same over the years, the institute maintains a "genetic stability program," for which it identifies mice with unusual phenotypes and removes them from the breeding colony.

"But those are actually interesting mice in and of themselves because of the mutations they may harbor," said Laura Reinholdt, a senior research scientist at the Jackson Laboratory in Bar Harbor, Maine, and the senior author of the publication. New sequencing technologies, she said, allow researchers to pinpoint these mutations much more quickly and at lower cost than in the past.

Four years ago, Reinholdt and her colleagues published a proof-of-concept paper showing they can capture and sequence the mouse exome. For their recent study, they sequenced the exomes of 172 strains of spontaneous mutant mice representing a variety of Mendelian disorders, including defects in behavior or neurological function, skin, growth or size, lifespan, craniofacial development, and skeletal morphology.

Through breeding, the Jackson researchers had already determined that the defects observed in these strains are heritable, and for most of them, they were able to map the mutation to a specific chromosome prior to exome sequencing.

Using an optimized pipeline for mouse exome analysis and a database that integrates sample and variant data, the researchers identified likely pathogenic mutations in 78 of the 172 strains, or 45 percent, a success rate that Reinholdt said is similar to what has been reported for clinical exome sequencing. For almost all of the solved cases, they had mapping data available, which narrowed the number of candidate genes.

They then took a second look at the unsolved cases. For some of them, they knew the chromosomal location of the mutation and had a strong candidate gene based on the mouse phenotype. When they analyzed the exome read alignments more carefully, they found mutations involving these candidates, often large copy number variants or structural rearrangements.

In addition, they performed whole-genome sequencing on five of the unsolved strains and found causative mutations in four, each involving coding regions of a single gene. One of the mutations was a 2.4-kilobase deletion, another a 300-kilobase duplication, the third a 7-base insertion, and the fourth a SNP that exome sequencing had missed because of low coverage.

Overall, manual exome analysis or whole-genome sequencing found mutations in 13 additional strains, or 8 percent of the total, and most of these were structural in nature.

In total, the researchers identified mutations in 53 percent of the strains — 89 potential pathogenic mutations in 73 genes for 91 strains. Of these mutations, 11 percent were in novel genes not previously associated with a mouse phenotype, and 37 percent in mouse genes that had not previously been associated with a human Mendelian disease.

Since submitting their paper, the researchers have performed whole-genome sequencing in more than 40 of the unsolved strains and found potential pathogenic mutations in about 60 percent of them, Reinholdt said. Her goal is to sequence the genomes of all the remaining unsolved exome cases if funding becomes available.

Even though most of the structural mutations found in the whole-genome data involved coding regions, these are tricky to find in exome data, especially insertions and duplications, Reinholdt said. Analysis tools to identify structural variants in exomes do exist, but she said her group has had "mixed results" using them. "I don’t think we have a perfect tool yet for that, and whole-genome sequencing, I think, is the way to go to find these," she said, though costs are higher than for exome sequencing.

For a fraction of the strains, neither exome nor whole-genome sequencing identified a putative causative mutation. Several reasons are possible: the mutations may reside in a poorly annotated region of the genome, which Reinholdt said is becoming less and less likely; the type of mutation may be too complex; or it may result from the translocation of a transposable element, which she said can be hard to find even in whole-genome data.

From mice to men

The results suggest it might be a good idea to follow up unsolved human exome sequencing cases with whole-genome sequencing. "I've always thought about exome sequencing as being able to catch the low-hanging fruit, but I think the gold standard is going to be whole-genome sequencing," Reinholdt said. "The limitation is the expense," not only for generating the sequence data but for managing and analyzing it.

Indeed, few unsolved clinical cases move from exome to whole-genome sequencing, except as part of a research study, according to Madhuri Hegde, executive director of the Emory Genetics Laboratory. Reimbursement for clinical exome sequencing "is not great" at the moment, and although CPT codes recently became available for both whole-exome and whole-genome sequencing, "it is a long way before we will see decent reimbursement," she said.

 Also, she said, many cases referred for diagnostic exome sequencing already had a CGH microarray test, which covers many structural variants, such as microdeletions or duplications, that whole-genome sequencing would detect, although the Jackson researchers noted in their paper that many of the deletions, duplications, and insertions they found were below the detection limit of commercial arrays.

Another way the mouse results could impact human disease diagnostics is by providing evidence for genes being disease-causing.

To see whether their results can inform unsolved human exome cases, the Jackson researchers used the GeneMatcher database to compare their gene list with candidate genes from unsolved human exome sequencing projects. They obtained a match for one gene, which had been found to be mutated in a family with a similar phenotype of seizures as the mouse model. They can now generate the same mutation found in the patients in mice and compare them directly. "That's an exciting direction for us — we have this really nice way to do comparative Mendelian genomics that we were not able to do very easily in the past," Reinholdt said.

Hegde cautioned that one needs to "exercise tremendous caution when applying [results from animal models] to diagnostic testing and interpretation," especially in the case of missense, silent, or splicing mutations. She said she has seen examples of genes identified in animal models where the phenotype did not match that of human patients, and it was difficult to interpret them clinically.

However, she said, it will be important to follow up on the genes and variants reported in the Jackson Lab study, and her group plans to include their results in the next version of its medical exome.