Q&A with Michael Orth Clinical neurologist Michael Orth, MD, PhD, knows the power of numbers. Big studies allow him to find so-called outliers—people with the HD gene who are much healthier or sicker than average, given the same number of CAG repeats (the expansion in the HD gene that causes the disease). To identify these extreme individuals, he taps into data from COHORT and REGISTRY, the studies that came before Enroll-HD. These outliers can teach us a lot, says Orth. Orth is a professor of neurology at the University of Ulm, where he also runs the HD clinic, and is science manager for the European Huntington’s Disease Network (EHDN). Q: Why are you interested in studying outliers? People with the same number of CAG repeats can be remarkably different. We think we can learn a lot from these differences. If we knew what protected them or made them worse, that might allow us to devise new approaches for therapy. Q: How do you use data from observational studies to find them? We can systematically look at a big group of controls [people without the HD gene] to see how cognitive performance and motor score evolve as people get older. For example, if you look at people carefully, you see that their eye movements change as they get older. On cognitive measures, performance drops as people get older. Education also plays a role—the higher the education, the higher the scores. All this needs to be taken into consideration. Now you know what the average is in the healthy population, and you can see what the influence of the disease is, beyond age, gender or educational levels. We look at people with a given number of CAG repeats and identify who is in the top 2.5 percent, doing really well compared with their peer group. And also who is in the bottom 2.5 percent. These are outliers or extremes. Q: What can we learn from outliers? One project I’m involved in is the genome-wide association study (GWAS) consortium, which uses DNA from about 4,000 people who took part in REGISTRY. We are drawing maps of the subtle differences in DNA that each of us harbors, and make us different without causing disease. We then ask if any of them are also associated with a particular phenotype (some observ- able aspect of behavior, health or thinking). By focusing on the extremes, you can take the analysis to a different level. If you focus on 100 extremes at one end and 100 at the other, you could ask: What differentiates those 100 who are doing well from those who are doing badly? Maybe you’d find a mutation in an important gene. Or you might look at the medications these people take, or their behaviors. Do they exercise regularly? What do they eat, what do they not do? That could lead to lifestyle advice. Q: Why do you need data from so many people? We use data from participants in REGISTRY and COHORT, and we hope to expand it with Enroll-HD. A unique feature of these studies is that they enroll thousands of people rather than hundreds. From a participant perspective, this may also mean that people have to be a bit patient. Recruiting these participants takes time, and working with the data also takes a lot of time. You have to look really closely at the data quality. I need to be confident that these people are outliers or ex- tremes not because their data is spurious, but because there is something to them that makes them special. But that way, you can ensure that your results hold water. Something based on results from 300 people may be a false positive. If the result were from 10,000 people, you can be fairly certain there’s something to it. This story was originally published in the September 2014 issue of Enroll!