Probing genes for disease risk

Alkes Price

December 5, 2014 — New research by Alkes Price, associate professor of statistical genetics at Harvard School of Public Health, and colleagues focuses on new approaches to characterizing and identifying genetic factors in complex disease.

What’s the basic finding of your new research?

Our study could help scientists determine the best way to search for genetic risk factors associated with disease. The study outlines a new way of estimating how much of variation in a particular trait is due to genetics. Examples we looked at in our paper include height, body mass index, and prostate cancer. Until about five years ago, the amount of variation in diseases or traits that is determined by genetic variation—called heritability—had been based on studies of twins, which look at differences in disease or trait values among identical vs. fraternal twins. In our new study, we describe an approach that relies instead on populations of mixed ancestry called “admixed” populations—in this case, African Americans, who inherit ancestry from both African and European ancestors. It’s useful to study these populations because—even though genetic differences between populations are generally much smaller than genetic differences within populations—the genomes of different continental populations contain valuable clues that can help geneticists determine the best way to search for genetic disease risk factors.

Why is the new finding important?

According to the twin studies, genetic factors account for 80% of some complex traits such as height—which is rather high. But geneticists have been unable to discover most of the genetic variants that are believed to play a role in particular traits and diseases. This gap is called the “missing heritability.” An important 2010 paper looked at a group of nearly 1 million genetic variants called “SNPs,” or single nucleotide polymorphisms. The human genome contains 3 billion letters in total; the 1 million SNPs in the study were a small subset of the more than 30 million SNPs known to vary among individuals. The 2010 study found that this small subset of variants accounts for roughly 45% of variation in height. But there’s a huge gap between 45% and 80%—what on earth is the source of that gap? Our new paper, looking at the problem by analyzing segments of the genome of distinct continental ancestry in admixed populations, found that genetic factors account for only 55% of variation in height, not 80% as previously believed. This makes the gap (now 45% vs. 55%) much smaller. So this finding underscores the importance of the 2010 paper, and suggests to researchers going forward that continuing to focus on those roughly 1 million common SNPs is likely a good strategy.

How could unearthing better information about genetic traits help researchers find new ways to fight diseases?

Ultimately, scientists are interested in disease traits. Geneticists tend to analyze traits like height and BMI because it’s very convenient to assemble huge amounts of data on those traits, because most people have their height and BMI measured. Studies of those traits can help inform studies of disease traits, for which there is often less data. The long-term goal is to identify disease-related genetic associations and then to use that knowledge to understand which genes or pathways are biologically important so that experimental drug targets can be devised, evaluated, and tested. The hope is that, eventually, drugs can be designed that can help cure or control disease.

— Karen Feldscher

photo: Aubrey LaMedica