Duke University's Nicholas Katsanis is taking a different approach to the search for disease genes: he is looking across related diseases for mutations and following up on those hits with functional assays. He has found that while one mutation may drive a disease, others can modify it.
"The traditional paradigm of disease gene identification is you have your candidate gene, you sequence some cohort of patients, and then you find some mutations and you take these patients off your test plate and you look for the gene for the next set of patients," Katsanis says. "We decided that there was no rational reason to do this."
Katsanis and his team instead looked across the entire patient population for mutations. They found that many patients had a primary mutation in one gene, but also had loss-of-function mutations in other genes. Those other mutations added to the patients' total mutation load, he says, and were linked to differences seen in the patients' clinical presentations. "The more dysfunction you carry, the worse off you are," Katsanis says. "You are always going to have primary sites that drive the majority of the disease, but then you are going to have additional variation that is of biological importance to whatever the disease gene is doing or fails to do."
In a recent Nature Genetics paper, Katsanis and his colleagues applied this approach to finding genes involved in ciliary dysfunction. "Let's start by taking genes that encode proteins that we know are necessary for ciliary biogenesis and function, and let's sequence them across a very large cohort that has many different flavors of ciliary disease," Katsanis says of his team's approach. In this paper, the researchers focused on the TTC21B gene — resequencing it in more than 700 cases and in nearly 400 controls — and then functionally assessed the variants they saw with in vivo and in vitro assays. TTC21B contributes mutations in about 5 percent of ciliopathies, the team reports.
"What you find is that different genes involved in ciliary biology contribute towards this mutational load of the entire system," Katsanis says. "Rather than sequencing one gene in one phenotype, we sequenced one gene in one biological system, mapping a range of clinical phenotypes that are known to map to the system."
Katsanis adds that he is concerned about what this interplay of genes means for all other ongoing sequencing projects. Right now, he says, researchers are not spending much time on interpreting sequencing data, opting rather to come back to it later. "I'm not sure we are equipped adequately to do this later," he says. "What is even more, we might be overhyping our interpretability."