Sifting through the genome

WASHINGTON D.C.—Aimed at discovering better way of searching among tens of millions of genomic variants to find those that make a difference in disease susceptibility and in other traits, the National Institutes of Health (NIH) has awarded six three-year grants in 2015 worth approximately $13 million, pending the availability of funds. The grants are administered by the National Human Genome Research Institute (NHGRI) and the National Cancer Institute, both parts of NIH.

As the popularity and potential of genomes rises within the research community, the possibilities for uncovering breakthrough treatments—and cures—become enticing.

The grants are aimed toward supporting research to develop new computational approaches for searching among millions of genomic variants to find the ones that really matter.

Comparing the genomes of many people suggests that there are tens of millions of genetic variants, or DNA spelling differences. For the last decade, scientists have used genome-wide association studies (GWAS) to find regions of the genome associated with diseases and traits.

In GWAS, the genomes of thousands of people with and without a disease are compared to find the genomic regions containing variants that affect disease risk. Although GWAS may find hundreds of variants that appear to be associated with a disease, it remains a challenge to find out which variants actually have a role in the disease process, and what that role might be.

“Before we can understand how a variant or gene functionally contributes to a disease—and then develop prevention and therapeutic strategies—we have to identify which genes and variants actually are involved in raising the risk,” states Dr. Lisa D. Brooks, program director of the NHGRI Genetic Variation Program. “We are looking for approaches that can find the causal variants out of the many variants associated with a disease, or at least narrow down the set.”

Most variants, including many that contribute to disease risk, response to drugs and traits such as height, are in genomic regions that do not code for proteins, Brooks says. These variants usually affect the regulation of genes, residing within “switches” in the genome that determine when and where proteins are made.

“If researchers can discover which variants cause the disease risk, then they can study the mechanisms, and hopefully develop methods or drugs to prevent or treat the disease,” she adds, noting that the issue has become more pressing as whole-genome sequencing becomes more common.

“For exomes, when variants were found in coding regions of genes, the genetic code provided a starting point for how to interpret the variation,” she says. “For example, researchers would first look for variants that changed an amino acid (non-synonymous differences), rather than ones that did not change an amino acid (synonymous differences).”

Brooks oversees the international 1000 Genomes Project, which aims to sequence the genomes of 2,500 people from 27 populations to find most human genetic variation, which is expected to serve as the basis for future studies mapping genes, genomic elements and variants affecting disease. She also manages the GWASeq project, which aims to sequence and characterize the variation in genomic regions associated with disease in about 4,000 to 8,000 people in each of five disease studies.

“We know a great deal about the protein-coding genes and what they do,” states Dr. Mike Pazin, program director in the NHGRI Functional Genomics Program. “For variants sitting outside the coding regions, it is difficult to know which parts of the genome they affect, let alone how the variants cause differences in function.”

“However, we know that 90 percent of associated variants found in GWAS are outside of the protein-coding areas,” Pazin adds. “Eventually, we want to understand mechanistically how the variants function in regulating genes, and how differences in the way they function affect disease risk.”

The researchers are developing computational approaches to combine many different sets of data to identify disease-causing variants or narrow down the set of candidate variants. They will use data from experiments to determine the accuracy of the computational predictions.

The following grants have been awarded (pending availability of funds):

To understand the DNA drivers of common human diseases (using immune diseases as test cases), the researchers plan to analyze how DNA variants associated with common immune diseases cause individuals to differ in their immune responses.

The researchers plan to interpret the importance of non-coding variants in human disease by studying their activity patterns and how variants influence chemical modifications called epigenomic marks, which affect gene regulation, and will develop statistical methods to identify locations in the genome where variants are more likely to affect the regulation of genes.

Although non-coding regions of the genome are not translated into proteins, they may be transcribed into RNA. Such RNA carries out various regulatory functions in a cell. The researchers have shown that disrupting RNA structure can lead to diseases in people, including an inherited eye cancer, retinoblastoma. They would like to develop computational approaches to predict structural changes in RNA that are caused by genetic variants.

The researchers will develop methods for interpreting non-coding genetic variation and for predicting disease-causing variants in the subjects’ genomes. They plan to develop various statistical models based on large amounts of information from individuals, and identify variants that contribute to hundreds of diseases and traits.

Previous studies have identified a number of DNA sequence variants strongly associated with age-related macular degeneration (AMD), which is the leading cause of blindness among seniors. In this project, the researchers will create computational models that can predict or narrow down non-coding sequence variants that contribute to the development of disease, using AMD as a test case.

The researchers seek to develop better ways to identify non-coding genetic variants that contribute to human disease. They plan to test a system they developed called Combined Annotation Dependent Depletion, which aims to identify which individual genetic variants contribute to disease. They will test and further refine the method as well as explore other related approaches in genome sequencing studies of both rare and common diseases.