Aproaches to studies of common & low frequency variants in complex human traits.
Oluf Borbye Pedersen
Application of exome sequencing on type II diabetes. Work in progress, since there is no final outcome to this exercise.
Genetic atlas for monogenic and polygenic traits. Allele frequency vs effect size. Rare alleles have high effect, low frequency = mendelian diseases. In other corner, common frequency, low effect = low penetrance diseases.
Diabetes, where is the missing heritability? Many genes cluster into the two corners above – many common alleles at low effect, only a few at high effect with low frequency.
Hyperclycaemia is a complex trait influenced by both genetics and metagenomics, as well as health behaviour. Many effects leading to same phenotype.
The cause of a geiven metabloic trait is unique in each individual. All of us have multiple risk alleles, genetic susceptibility involves many rare and common risk alleles.
“The responses to current preventive and therapeutic interventions differe substantially among individuals suggesting major heterogeneity.” – US FDA.
The future requires individualized diagnosis, prevention and treatment.
High density SNP array genotyping. Limitation: they only cover common variants, down to 5%, which isnøt enough to find the rare variants. At least 37 different loci associate with Type II diabetes. Each of them are common, but only increase risk 5-35%. However, they do cluster into several pathways. Obesity, insulin resistance, beta-cell dysfunction, reduced incretin effect, etc.
Explosion in the number of validated common variants for cardio+metabolic traits. Gene and causative mutations are often unknown. Indentified variants only explain 2-10% of genetic contribution.
We have taken the first steps from genomics to biology, but still have a LONG way to go before this will move into a clinical situation.
So, how do we fill in the gaps and find the rest of the heritabity and risk factors? Additional common risk alleles? Gene+environment interactions? Copy number variations? Epigenetics? Gene-gene interactions? Many places it could be hiding.
We have a long way to go.
Started on a project of 2000 Danes. 1000 of them have T2D, obese and hypertensive, 1000 Danish controls. Nimblegen 2.1 Human Exome array, illuina analyzer II with greater than 8x coverage.
Genetic association studies in LuCamp. Stage 1: exome sequencing. [BGI exome report stats – 56,594 in dbsnp, 13,488 novel snps] Stage 2: 20,000 snps on customized illumina array. 16,000 danes then screened on the array, covering a range of T2D, hypertension, etc.
Results were as expected [Graphs covering a lot of metrics being tossed up on the screen faster than I can take notes on.]
Selected type 2 diabetes variants from project have lower frequency…. [running out of time, so slides are going by way too quickly.]
Stage 3: of projects, Validation of best snps will now be tested on 30,000 europeans.