How does one discover a gene responsible for heart attack risk?

Myocardial infarction (MI) or heart attack is the leading cause of death in the U.S. MI is the result of interplay between genetics and lifestyle factors. Understanding the inherited basis for MI can pinpoint mechanisms that cause MI in humans and point to new avenues for therapy.

There are 3.2 billion letters of human DNA sequence.  Which specific letters are responsible for risk for MI?  Approaches to identify genes and DNA sequence variants that cause MI include:

  1. Linkage
  2. Association
    1. common variant association
    2. rare variant association

If MI segregates in a family with a pattern consistent with the ratios described by Mendel, linkage analysis and/or sequencing may be used to directly identify the gene and variant responsible for MI. There are several Mendelian syndromes where early MI is a prominent feature and nearly all of these involve high-levels of low-density lipoprotein cholesterol. See review by Rader DJ, Cohen J, Hobbs, HH.  J Clin Invest 2003.

In nearly all other cases of MI, association studies are the key approach. Association studies involve testing whether the frequency of a set of one or more alleles differs between cases and a control population, indicating that the set of alleles is associated with the disease.

Common variant association studies (CVAS). Common variants are those frequent enough that it is practical to test each variant individually by estimating its frequency in cases and controls.  An operational definition for common is >0.5% in frequency, corresponding to one heterozygous carrier per 100 people. Association studies of individual common variants are also referred to as genome-wide association studies (GWAS).

Rare variant association studies (RVAS). Rare variants are those <0.5% in frequency. RVAS differs from CVAS in two respects. First, rare variants are too numerous to catalogue comprehensively and so must be directly enumerated in every sample by DNA sequencing rather than by genotyping known variants. Second, and more fundamentally, rare variants occur too infrequently to allow association tests of individual variants. RVAS thus require aggregating rare variants into sets and comparing the aggregate frequency in cases vs. controls.