February 28th is International Rare Disease Day, and it reminded me that I’ve been rather sporadic in updating this blog. That’s mostly because I’ve been working on rare disease analyses, publications, and grants. However, it’s motivated me to share a bit about my strategy for studying rare diseases with genomics and how my thinking has evolved in the past year.
Rare Disease Research at NCH
Our research study on the genomic basis of rare diseases has been running for about three years. To date, we’ve enrolled more than 200 individuals from about 65 families. A wide range of phenotypes are represented, including congenital malformations, developmental disorders, neuromuscular diseases, and other conditions.
We’ve found a likely cause of disease or are pursuing a candidate variant in roughly 1/3 of the cases analyzed so far. Three of those were published last year, and several more are submitted or in preparation. This is consistent with the widely-established diagnostic yield (25-30%) for studies like ours. Our clinical laboratory has a similar diagnostic rate for WES cases.
It’s daunting sometimes. We want to provide an answer for every family, but statistically speaking, we’re more likely to have negative results. However, there is value in both success, i.e. providing an answer, and in failure. Here are some key lessons learned.
Rare diseases are more common than you think
This sounds counter-intuitive at first, but it’s something that is quite apparent when working at a major children’s hospital. Any particular rare disease affects only a small proportion of the population by definition. For example, consider Rubenstein-Taybi syndrome, a rare congenital malformation syndrome with an estimated prevalence of 1 in 100,000. That’s pretty rare. Even so, at least three patients have been diagnosed with RSTS in the past couple of years at our institution alone.
That’s just one example of a rare disease. At least 7,000 rare diseases have been described. Consider this infographic from NORD, the National Organization for Rare Disorders:
While individually rare, they collectively are a significant cause of medical conditions. Especially in children. Granted, there’s undoubtedly a selection bias: Kids with rare disorders are more likely to be hospitalized and be seen by a geneticist. Also, many genetic disorders cause chronic medical issues, which means we see them a lot more often than the average healthy kid. This probably makes them appear more prevalent than they really are. Even so, there are 30 million people affected by rare disorders in the United States alone. That’s about the same number as people with diabetes, one of the most common diseases.
Precise and detailed phenotyping is crucial for diagnostics and research
Many rare diseases, especially the well-studied genetic syndromes, have a “classic” clinical presentation. Clinical geneticists are taught to recognize patterns of human malformation that manifest in certain conditions. However, there’s a growing appreciation for the phenotypic spectrum of disease. In other words, not everyone with the same rare condition has the same set of medical problems (or characteristic features). Further, many of the genetic variants underlying rare disease have incomplete penetrance, i.e., not everyone with them gets the disease.
This is why precise and thorough phenotyping is critical in our field. I don’t just mean the patient, but also parents, siblings, and distant relatives. Genetic diseases tend to run in families, and the family history often guides our approach to analysis. When it comes to pedigrees, the word “normal” is a dangerous thing. Sometimes it means perfectly healthy, but other times it means they haven’t come to medical attention as far as we know. Although de novo mutations account for a significant proportion of genetic diagnoses, many causal variants turn out to have been inherited from “normal” parents.
Data sharing and collaboration are keys to success
This is especially true for rare disorders because the patients are usually few and far between.
Back in September, I heard a wonderful talk by Dr. Kym Boycott of the Children’s Hospital of Eastern Ontario who is one of the leaders of Canada’s Care4Rare Consortium. That consortium has provided a genetic diagnosis for 1,500 patients and families, and discovered 135 new disease genes along the way. One of their keys to success? Data sharing. They share genetic and phenotypic data with other investigators via the Matchmaker Exchange (a free service that unifies GeneMatcher, MyGene2, DECIPHER, and other data sharing portals). Further, they make their sequence data available to any qualified researchers who want to have it. They share everything, which I suppose is not surprising since they’re Canadian.
Dr. Boycott’s talk motivated me to start sharing our rare disease research data more systematically. We submit candidate variants/genes of interest to Matchmaker Exchange through GeneMatcher, and we submit variant interpretations to the ClinVar database. These efforts have already paid off; at least three of our rare disease cases now have a diagnosis because of data sharing efforts.
With new genomics tools & approaches, we’re better equipped than ever to study rare diseases. Even so, sharing and collaboration are essential for the discovery and characterization of new disease genes. There are millions of people around the world counting on us. Let’s get to work.