Go Back

Back

The earlier forms of predictive medicine

Next

Genome-wide Association Studies

First things first
Genome-wide association studies rely on huge amounts of data. Algorithms comb through this data searching for differences in the genetic makeup of individuals. Then the algorithm can find the specific single nucleotide polymorphisms (SNPs) that can be correlated with a medical disease.

GWAS

Step 1
First, we filter through the data and remove SNPs that don't represent significant variability or have large amounts of missing data.

GWAS

Step 2
Second, we filter through the data and ignore certain individuals for various reasons. Either we may not have enough data, or the sample could have been contaminated.

GWAS

Step 3
Third, we impute non-typed (non-genotyped) SNPs to our data. Current genotyping systems cannot geneotype all SNPs so we must use external sources. External projects such as the HapMap Project have created haplotypic maps. With these maps, we can predict non-typed SNPs using information we already have about typed SNPs.

GWAS

Step 4
Fourth, we can determine the association of each SNP with the medical disease. This allows us to identify candidate genes that correspond with the disease.

GWAS

Step 5
Lastly, we test our candidate genes further to establish a direct connection between the SNP and the medical disease. Thus, we can efficiently and (sort of) easily determine the risk of an individual contracting a certain disease with only his or her genetic makeup.