Genome-wide association study (GWAS)

John Timmer – May 17, 2008 7:25 pm | 0

A genome-wide association study (GWAS) is an attempt to identify weak genetic influences on a trait that is influenced by multiple genes. It relies on scanning the entire genome in a large number of individuals in order to provide a statistically robust identification of genetic factors that, separately, have an influence that is quite small. Unfortunately, the large number of potential associations tested in a single study, as well as differences in the methods of testing, have produced some specious results.

Typically, testing is done using a DNA chip that is designed to identify common variations in the human genome, like single base pair differences. Each chip can contain thousands of these variants, spread evenly throughout the genome. Thus, putting an individuals DNA on one of these chips can be sufficient to provide an individual’s genotype throughout their entire genome. For a single individual, it’s impossible to identify which of the thousands of individual variants are associated with a disease. By testing a large pool of individuals, however, it is possible to identify some areas of the genome that are more commonly associated with a trait, such as height or susceptibility to diabetes, even if the influence of individual factors is small.

To provide an accurate genotyping, however, a DNA chip must contain thousands of individual variants. The standard statistical measure for accepting a hypothesis in biology is a five percent probability of the association occurring by chance. Thus, at that probability, there can be dozens of spurious associations in a single experiment. There are ways of performing more rigorous statistical analyses, but they were not always done in early GWAS studies.

Additional problems can arise from the fact that DNA chips are made by a number of manufacturers, and there are differences in the variants covered, as well as the experimental use of these different chips. Thus, an initial finding of variants associated with human height was called into question because of the specific chips used in the work.