We propose an analysis of the prevalence and health consequences of genetic variants that are not single nucleotide polymorphisms (SNPs). These include deletions, duplications, variations in copy number and translocations of parts of the genome, up to and including the duplication or deletion of whole chromosomes (monosomy, trisomy etc). Specific examples are: duplications/deletions of parts of genes such as the Haptoglobin duplicon (HP gene) and growth hormone receptor exon 3 insertion/deletion (GHR gene), large chromosomal alterations such as Prader-Willi syndrome (deletion of part of chromosome 15) and changes in chromosome number such as Klinefelter's syndrome (XXY, an additional sex chromsome), triple X syndrome (XXX, an additional sex chromosome) and Down's syndrome (three copies of chromsome 21). These types of genetic variation are not directly measured by current genome-wide SNP arrays, but can be inferred from the raw data, and present the opportunity to gain added value from existing data.

Experimental plan:

* Existing raw genotyping data from Illumina SNP genotyping performed at the Wellcome Trust Sanger Institute will be used

* Raw fluorescence data will be analysed for reductions or increases in total fluorscence across a number of genomically adjacent SNP tests (representing loss or gain of the DNA sequence containing those SNPs

* "Extended homozygosity" (regions where many SNPs appear to have two copies of one allele, rather than two different alleles) will also be identified - some of these will represent deletions

* Pairwise Hardy-Weinberg equilibrium will be used to identify distortions in genotype frequencies that might represent a "null allele" in the sample

* All identified variants will be "phenome-scanned" (i.e. tested against a range of phenotypes representative of all data available), and/or specifically analysed (if they affect a single gene for which the function is known). Whilst numbers may be small, the aim is to identify whether a genetic variant causes individuals to be in the "tail" of the population distribution for particular phenotypes

* Population frequencies of variants, and potentially differences in break-points, will be estimated. Further testing of DNA for breakpoint location and fine mapping is beyond the scope of the current application, but may be requested in the future

For phenotypic analyses a representative range of phenotypic data will be needed. We ask the committee to approve "all" phenotypes, as genomic analyses will inform the phenotypes to analyse on a case-by-case basis, and as such phenotype choice will be iterative. In each case we would wish to compare relevant phenotypes for those individuals with a genetic variant to the phenotype distribution or frequency in the rest of the ALSPAC cohort.