GWAS: scanning the genome for associations
A genome-wide association study (GWAS) takes a large group — say, thousands of people with a condition and thousands without — and compares millions of SNPs across both groups. For each SNP it asks one statistical question: is a particular letter more common in one group than the other? SNPs that are reliably more frequent in cases are flagged as *associated* with the trait.
GWAS works best for complex traits — height, blood pressure, diabetes risk — where no single gene rules. Each associated SNP nudges the odds a tiny bit; hundreds of them together do the steering. Adding up someone's risk-raising variants gives a polygenic score, a rough estimate of inherited predisposition. These scores are statistical tendencies across populations, not predictions about any one person's fate.
GWAS in one picture (one SNP among millions):
has 'G' has 'A'
Cases 62% 38%
Controls 41% 59%
^ 'G' enriched in cases
-> SNP flagged as ASSOCIATED with the trait
Reminder: association != cause.
The 'G' may just ride along with the true variant
nearby (its haplotype), and effect size is small.So what does “having your genome sequenced” mean?
Having your genome sequenced means a machine reads your ~3.2 billion letters and bioinformatics compares them to the reference, producing a list of the places where you differ. That list — your personal set of variants — is the real product. Most of those differences are shared with millions of others and mean nothing on their own.
- Your DNA is sequenced, aligned to the reference, and reduced to a list of variants where you differ.
- Each variant is annotated: is it common or rare, in a gene or not, harmless or known to matter?
- A few may be clinically meaningful; many are uncertain; an incidental finding may surface something you weren't looking for.
- Results describe probabilities and tendencies, not certainties — and they are best read with a professional.