A multi-ancestry genome-wide association study (GWAS), led by scientists from the Global Lipids Genetics Consortium (GLGC) recently analyzed genomic data from nearly 1.65 million individuals to narrow down the number of genomic variants strongly associated with blood lipid levels and compute more predictive polygenic risk scores for heart disease. Increased levels of lipids such as cholesterol, triglycerides, and lipoproteins in the blood pose a heritable risk factor for heart disease.
The study, funded by the National Heart, Lung and Blood Institute (NHLBI), was published in the journal Nature on December 8, in an article titled, “The power of genetic diversity in genome-wide association studies of lipids.”
Cashell Jaquish, Ph.D., a genetic epidemiologist, and program officer in the division of cardiovascular sciences at the NHLBI said, “These results show that our concerted effort to include many diverse groups of people in genomic research will yield benefits such as new therapeutics and prevention strategies that improve the health of all people.”
Within the estimated six billion-base human genome, are sprinkled four to five million genomic variants that either increase or decrease our risk for a disease or have no effect at all. To determine how genomic variants influence the risk for specific diseases, researchers compute polygenic risk scores that provide an estimate of an individual’s risk for specific diseases, based on their DNA changes relative to the total number of variants linked to a disease.
“Finding the set of genomic variants that are important for this trait is key for us to understand the biology and identify new drug targets,” said Cristen Willer, Ph.D., senior author and professor of human genetics at the University of Michigan, Ann Arbor. “These genomic variants then inform how well polygenic risk scores work to determine risk for such diseases.”
Although nearly 6,000 genome-wide studies have focused on the association of specific genomic variants and cardiovascular disease, most of these studies have included participants solely from European ancestries.
Amy Bentley, Ph.D., NHGRI staff scientist at the Center for Research on Genomics and Global Health and a co-author of the study said, “Polygenic risk scores are gaining prominence as a having potential clinical utility to identify individuals at increased genetic disease risk. Unfortunately, polygenic risk scores rely on large, genomic studies for their development, and as a result of the well-recognized underrepresentation of individuals of diverse ancestries in genomic research, the performance of polygenic risk scores is much better for individuals of European ancestry compared to other ancestry groups.”
“This study includes a large number of non-European ancestry individuals, allowing for the development of polygenic risk scores that perform well across populations, an important advance in efforts for equity in genomics,” Bentley added.
To address this lack of genetic diversity, GLGC scientists collected GWAS data from 1,654,960 individuals in 201 primary studies. These individuals represented five ancestral groups with 6% African, 8.9% East Asian, 79.8% European, 2.9% Hispanic, and 2.5% South Asian. About 350,000 participants in the study were from non-European ancestries.
The primary GWAS studies contained data on blood levels of five blood lipid traits including low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), triglycerides (TG), total cholesterol (TC), and non-high-density lipoprotein cholesterol (nonHDL-C).
The researchers computed polygenic risk scores based on data from each of the different ancestral groups, separately and combined. They tested the risk scores in a diverse set of studies including the Africa America Diabetes Mellitus (AADM) Study that enrolled participants from Ghana, Kenya, and Nigeria. Charles Rotimi, PhD, scientific director of the NHGRI Intramural Research Program, was the principal investigator of the study. (For more on genomic research in Africa and transcontinental collaborations read this GEN Insight article).
The meta-analysis for blood lipid traits confirms that contributions of common genetic variation to blood lipids levels are largely similar across diverse populations.
The polygenic scores computed from the diverse genomic data also showed that the inclusion of genetic diversity in the dataset results in scores that are much more predictive of whether a person of any ancestry will have elevated low-density lipoprotein cholesterol levels than a score that only includes European genomic data. For each ancestral group, the polygenic risk score that used data from all ancestries worked equally well or better than the risk scores derived from data from the same ancestral group.
“The message couldn’t be more clear. To have a fuller understanding of the effects of genomic variation on disease, we simply must include as many diverse groups of people as possible,” said Rotimi, a co-author on the paper. “It is the single biggest way by which we can ensure that the gains of genomic medicine and technologies are equitably deployed to serve the health needs of all human populations.”
Although genomic studies of blood lipid levels have helped generate critical mechanistic insights and druggable targets for cardiovascular diseases, the lack of diversity in the genetic ancestries of the participants in earlier studies may have missed genetic variants that contribute to lipid level variation in other ancestries due to differences in allele frequencies, effect sizes, and linkage-disequilibrium patterns. The current large-scale study underscores the importance of including diverse groups of people in genomic studies to achieve a more comprehensive understanding of the effects of genomic variation on diseases. This will facilitate the application of insights from genetic variants in preventive and precision clinical practices.
Two additional projects from the Global Lipids Genetics Consortium have been pursued concomitantly and will be published soon, said Bentley. These studies that use data from the multi-ancestry GWAS of blood lipids to perform pheWAS [Phenome Wide Association Study] probe into the effects of genetically predicted lipids on other diseases. The studies aim to prioritize putative causal genes using a variety of gene prediction methods, and to identify mechanisms that regulate plasma lipid levels by integrating new and existing epigenomics datasets.