Quantitative Trait Locus (QTL) Identification in Isolate Populations

 

Summary

The aims of the study are (i) to genetically map and identify quantitative trait loci (QTL) concerned with common human disorders in isolated populations; (ii) to characterise the genetic architecture of complex traits; (iii) to evaluate the use of genetic markers in trait prediction and treatment.

 

 

 

 

 

 

Aim 2: The dense marker data will also be used to estimate individual genome-wide heterozygosity which is being analysed in relation to trait values to test the hypothesis that these are, in some cases, correlated. This type of analysis provides information on which traits have dominance components, and on the number of loci involved (see Rudan et al .2006).

 

Aim 3 We are currently investigating the predictive value of genetic markers associated with complex traits.

 

Figure 2. Association of SLC2A9 SNPs (-log10 P-values) with serum uric acid in 986 Croatian islanders. following a genome wide association scan using 317,503 SNPs. The location of SNPs are shown across a 700 kb region on human chromosome 4p16.1 containing the SLC2A9 gene. The genomic location of exons in the long and short SLC2A9 isoforms are shown.

The genome-wide significance threshold is shown as a red horizontal line. Below are the linkage disequilibrium patterns of the Hapmap dataset within 240kb centered on the top SNP, using Haploview

 

Current and Future Work

Common variants identified in our own and many other GWAS studies have, in general, been of small effect and have not explained much of the total genetic variance for the traits examined (usually <10%). Efforts are now underway to examine the influence of rare variants on quantitative traits, many of which are predicted to have larger effect sizes, using a variety of methods including whole genome exome sequencing and haplotype analyses. Isolated populations have a potential advantage for this next phase of the genome project due to the availability of faster and deeper sequence coverage provided by next generation sequencing.

 

Isolates are potentially advantageous because the degree of relationship between random individuals sampled from these populations is higher than with samples drawn from large urban ones, allowing us to infer genotypes identified by exome sequencing in a small subset of the population (e.g. in 10%), in a high proportion of the remaining population who have not been sequenced (but have had whole genome SNP genotyping). This approach is currently under investigation. We are also following up confirmed GWAS associations by attempting to identify causal variants using both bioinformatic and laboratory based methods. Finally, we are performing expression and functional assays to identify the functional consequences of identified associations at a molecular level.

 

 

Lab Members

Current lab members involved in this work are:

• Professor Nicholas Hastie (joint P.I.)
• Professor Alan Wright (joint P.I.)
• Caroline Hayward PhD (MRC core staff)
• Veronique Vitart PhD (MRC core staff)
• Pau Navarro PhD (grant funded)
• Susan Campbell BSc (MRC core staff)
• Jennifer Huffman BSc (PhD studentship)
• Chloe Stanton (MRC core staff)
  

 

1. Quantitative Trait Locus (QTL) Identification in Isolate Populations
   (this page)
2. Quantitative Trait Locus (QTL) Identification in Isolate Populations Publications

 

 

Approach, Progress and Future Work

 

Aim 1: In order to identify QTL influencing disease-related quantitative traits (QTs), we have performed genome-wide linkage and association analyses in samples obtained from an epidemiological field survey of over 2,000 adult volunteers from the islands of Vis and Korcula on the Dalmatian coast of Croatia (Figure 1), in collaboration with colleagues at the Centre for Global Health, University of Split (P.I. Prof. I. Rudan) , the University of Edinburgh (P.I. Prof. Harry Campbell) and elsewhere. The advantages of studying such populations include the high rates of participation, relatively uniform environmental, suitable genetic characteristics of the population (e.g. increased relatedness) and ability to recruit families. We are working closely with colleagues at the University of Edinburgh on a similar study in the islands of Orkney (P.I.s Dr. J. Wilson, Prof. H. Campbell)) (Figure 1).

 

We have measured >300 biomedically relevant quantitative traits in all members of the Vis and Korcula isolates, and estimated the genetic parameters influencing each trait. We then carried out a 320,000 single nucleotide polymorphism (SNP) genome-wide association study (GWAS). QT association studies were then carried out which identified a number of strong associations including ones affecting serum uric acid and gout (Vitart et al., 2008), lipid levels and coronary heart disease (Aulchenko et al., 2009) and height (Johansson et al., 2008). Many other associations are currently being followed up or have been published (see Key Publications list). The association of SNPs within the SLC2A9 gene with serum uric acid proved particularly interesting since we were able to show using oocyte transport assays that the encoded protein, GLUT9, previously known to be a hexose transporter, was a strong uric acid transporter (Vitart et al., 2008).

 

Minor Allele Frequency Distribution

Figure 1. Allele frequencies differ between Croatian island and mainland European (CEU) populations. There is an excess of low frequency alleles in the isolated population (Navarro et al., 2010).

 

Professor Rudan and colleagues have extended the field work to include a further 2000 individuals from the Croatian mainland (Split), which will allow comparisons between urban and island population characteristics. We are also members of the EU FP6 EUROSPAN consortium of isolated European populations (Sweden, Netherlands, Croatia, Scotland, Italy), in whom similar traits have been measured and GWAS scans performed. This has identified a number of genetic associations (e.g. Johansson et al., 2008) and allows us to compare and contrast the results in different populations, who often show contrasting environmental influences (Figure 2).