Lecture 9 Flashcards
Why is genetics a mathematical problem?
Most diseases that have a major economic, social and health burden have a polygenic component.
Need to use statistical techniques to identify the genes (and the environmental factors) involved in the disease.
What two steps are needed to calculate a polygenic risk score?
- Discovery GWAS
- Target study
- Independent of discovery study
- Construct polygenic risk scores (PRS) for individuals
PRS = calculated by the sum of an individual’s risk alleles, weighted by risk allele effect sizes derived from genome-wide associated study (GWAS) data
Provides individual-level measures of genetic loading for disorder or trait
Genetic liability to disease is approximately normally distributed
True or false
True
How useful is this polygenic risk score for
individual level risk?
And for population level assessment?
Not very – prediction is currently poor
Better at population level to distinguish between cases and control
- According to a study by BMJ investigating type 2 diabetes, how well do genes and environmental/clinical risk factors predict individuals who develop type 2 diabetes?
- Did this study calculate the polygenic risk score?
- How did a iCoGs editorial, Nature Genetics 2013 differ to the above study? What was found concerning the genetic contribution to study diabetes?
- Clinical & family history risk factors provided good prediction (red line)
Adding genetics to this did not improve prediction
Why not? - main reason is that only a small component of the genetic contribution to type 2 diabetes is identified - 20 SNPs investigated.
- No. Calculated gene count score rather Polygenetic risk score, to do so added together the number of risk alleles that individual carried not weighted by effect size
- Larger number snps studied here, however this was found to explain only a small aspect of genetic contribution to diabetes
SNPs may be significantly associated with a disease in genome-wide association studies, but provide little information to assess disease risk accurately.
Why?
- Incomplete information: more SNPs to be detected with larger sample sizes
- Rare variants often not tested for association
- SNPs detected are not causal variants : tagging SNPs, in LD with causal variants
- Environmental risk factors not assessed
- Gene-gene or gene-environment interactions may be important
Why should polygenic risk scores be used/considered to develop a more effective breast cancer screening programme?
Most women reach 2.5% 10-year risk between age 45 and 50 years - aligns age screening is offering
However, women with higher polygenic risk score have increased risk 10-20 years earlier than when mammogram screening is offered and therefore are not screened
Some women never reach this level (no screening?)
What did Khera et al. 2016 find in concern to Coronary artery disease?
Genetic and environmental risk factors work together to predict coronary disease
Cumulative effect across both sources
These risks are not predictive enough for an individual’s risk, but provide useful population level information -
Low genetic risk: Poor environment will still increase risk
High genetic risk: Can protect against CAD by ensuring healthy environment
What are the clinical concerns of direct to consumer genetic testing?
Misinterpretation of results
Results have no clinical implications
Increased rates of anxiety
Demand for inappropriate follow-up medical tests
Should you find out you are a BRCA1 mutation carrier online?
- Very limited individual prediction available from current findings due to what factors?
- What is prediction even weaker for?
- What might this be valuable for?
- Incomplete knowledge of polygenic component of disease
Causal genetic variants are unknown
- Prediction is even weaker in non-European ancestry populations
- Might be valuable for developing screening programmes
Better individual prediction is currently derived from…
Family history
Environmental risk factors (smoking, body mass index)
Pre-clinical factors (blood pressure, cholesterol levels)
