Trigger 10: Mendelian randomisation Flashcards
mendelian randomisation use
previously collected data e.g. GWAS (biobanks)
mendelian randomisation simple
used to strengthen causal inference
epidemiologists are interesting in understandings
factors which relate to health and disease e.g. smokers dye younger than non-smokers
main issue relates to
determining cause and effect due to confounders e.g. smokers are more likely to drink more and have less healthy diets
an RCT where 50,000 people are randomised to smoke ad 50,000 are randomised to not smoke
would prove a cause and effect relation - however would not be ethical or practical
our genes influence how much
we eat, drink, smoke and eat
genetic influences are not effected by
Confounding factors –> anything you may choose to do or not to do e.g. smoke
mendelian randomisation study can tell us
about which factors cause disease
example of MR study
heavy smokers may carry specific genes and die younger than the average individual, but mendelian randomisation also show that non-smokers with the same genes have normal life expectancy - showing that smoking causes shorter life expectancy
MR uses a ……. approach
statistical approach, which lies to determine if exposure to certain factors causes certain diseases - unbiased estimate
example of cause and effect with confounders
e.g. if you looked at correlations you may see that people who drink more get lung cancer, suggesting that alcohol causes lung cancer. - however this is confounded by the fact that those who drink more also smoke more - therefore smoking is causing individuals to drink more - smoking could also be causing cancer
draw flow chart which shows how cause and effect can be confounded
to find genes which cause common traits what sort of studies are used
genetic association studies e.g. MR
genetic association studies correlate
commo genetic variants (SNPs) with disease prevalence
geen variants are present
more frequently in people with disease than control- providing proabbility of disease status
which statistical tests can be used to see if there is a statistical difference between cases and controls relating to specific genotypes
Chi-sq
