The second was an advanced and detailed approach to epidemiology. Most genetic studies seek to capture as large a segment of the population as possible, which can be costly and can favor the number of subjects over the quality of data collected from them. Hobbs and Cohen’s proposal instead opted to direct resources toward creating the most detailed subject profiles possible.
“We decided that for everything that we measured, we would use the best, most accurate assay available,” Hobbs says. “We would have very precise phenotype traits. When we looked at the heart, we would use magnetic resonance imaging, the very best imaging modality of the heart. When we looked at fat in the liver, we didn’t use sonography, but rather another methodology called proton magnetic resonance spectroscopy that is much more accurate at measuring liver fat.”
When Hobbs decided she wanted to use an electron beam computerized tomography scanner that UT Southwestern didn’t have, she phoned the dean and asked him to commit to purchasing the million-dollar machine if they got the grant. He agreed. The final grant proposal told a compelling story of a medical institution coming together around a population-based genetic study of heart disease that would capitalize on its genetically diverse local gene pool and invest in collecting the most detailed medical assessments of any study of its kind. Hobbs and Cohen, joined by other colleagues from Dallas, flew to Las Vegas to present their proposal in the final round of considerations. They won.
“We were more surprised than anybody,” she says, “though we thought we had written a really great grant.”
Over the next three years, the team sent 50 people out into the field to knock on doors and ask questions about the family histories of random Dallas residents. They narrowed down which household member might make the best candidate for the study, and when the subjects were chosen, they underwent an array of tests. By 2003, Hobbs and Cohen and a team of other scientists at UT Southwestern had amassed a database of detailed medical profiles on a small but critical cross section of the population. They began analyzing their data by sorting their subjects into phenotypes and looking for extremes in the distribution of various traits. The first trait they looked at was high-density lipoprotein (HDL) cholesterol levels, isolating individuals with the highest and lowest HDL levels and sequencing the genes in the hope of finding shared mutations.
“We were asking whether there might be sequence variations that cause low HDL, not just in individuals with rare genetic diseases but in the general population,” Hobbs says. “And we found that it was the case. We found that healthy people were riddled with mutations.”
That’s when a team of researchers in France discovered several French families with hypercholesterolemia, a condition of extraordinarily high cholesterol, who all shared a similar genetic mutation. The French researchers isolated the gene, which produces a protein called proprotein convertase subtilisin/kexin type 9, or PCSK9. When they published their research in 2003, Hobbs wondered if she could find similar connections between PCSK9 and cholesterol in the participants in the Dallas Heart Study. If the French researchers found that a high level of PCSK9 is related to high cholesterol, Hobbs wondered, were any of the participants in the Dallas Heart Study benefiting from the opposite effect? Was there a mutated form of PCSK9 that could result in low cholesterol?