e propose to conduct a gene x gene and gene x environment analysis of candidate factors that have been shown to increase susceptibility for dyslexia and SLI in previous research or that have been chosen on the basis of their biological function. We will genotype all the available ALSPAC children DNA samples (greater than 10,500) for susceptibility variants within candidate genes and conduct the analysis using cognitive outcomes and environmental exposures already assessed in the ALSPAC project.

The genotyping will be performed using the Sequenom i-plex technology in collaboration with the Genomics Core facility at WTCHG. We already have in our laboratory the sufficient quantity of ALSPAC DNA required by this project.

We will select for analysis measures of reading, language and speech abilities, including measures of single word reading, non-word reading, phoneme awareness, orthographical skills, spelling, short-term memory (including NWR), verbal expression, comprehension and IQ. Many of these measures are the same as those used in our current dyslexia and SLI linkage and association studies. We plan to include in the analysis measures of attention and hyperactivity behaviour since attention deficit hyperactivity disorder (ADHD) shows extensive co-morbidity with both dyslexia and SLI.

The environmental factors will include measures that have already been proposed as risk factors for SLI, for example otitis media with effusion (OME) and maternal education, or associated with dyslexia, for example the home literacy environment and Omega-3 fatty acid consumption. We will consider also the child's environment, more general aspects of parenting and parent-child interaction. Statistical analysis will be based on linear regression models extended to include one or more covariates. We will also interrogate for multilocus interactions and genetic effects specific to factors underlying the phenotypes using techniques such as principal component or cluster analysis. Analysis will be performed both for single SNPs and haplotypes.

A detailed list of genotypes, phenotypes and environmental measures is given in the appendix.

Strategy details

Aim 1: To investigate whether the effect of dyslexia and SLI susceptibility variants can be modulated by different genetic factors (gene x gene interactions). We will select the maximum number of SNPs that can be accommodated in three i-plex reactions. Most of the novel dyslexia candidate genes we are proposing for this project have been selected for showing a strong signal in the recent WGA analysis we conducted in 600 individuals with dyslexia collected in Germany and in the UK as part of the collaborative NeuroDys project (www.neurodys.com). Other genes have been selected for their (i) similarity to other established dyslexia candidates (KIAA0319-like), (ii) interaction with established dyslexia candidates (we have recently identified interaction between the AP2M1 and KIAA0319 proteins) or (iii) role in neuronal migration, a biological pathway to which most of the reported dyslexia candidates seem to contribute. We will also include candidate genes for ADHD since this disorder show an extensive co-morbidity with dyslexia and SLI. SNPs have been selected either for showing significant association in previous work or for tagging the most common haplotype across the genes of interest. Statistical analysis will be conducted at different levels. We will first test for association between single SNPs and quantitative measures of reading, language and speech in a linear regression framework. This analysis will enable us to prioritise further investigations on the basis of the strength of association found with each gene. Following these initial association analyses, we plan to test for gene x gene interactions between the genotyped loci for example by integrating an epistatic component into the regression model. In this second step of the analysis we will incorporate also genes for which we have already obtained ALSPAC approval*.

A specific question we aim to answer is whether any tested SNP or haplotype can modulate the effect of the KIAA0319 risk haplotype (which we have recently identified within the ALSPAC sample) and whether it is possible to detect any specific interactions between candidate genes involved in the same biological pathways, such as between genes involved in neuronal migration.

Aim 2: To investigate the role of environmental factors in modulating the effect of genes that contribute to measures of reading, speech and language (gene x environment interactions). This analysis will aim to test whether any environmental factors can modify the effect of an individual's genetic background in either a protective or adverse manner for reading and language ability. The SNPs, haplotypes and gene x gene interactions that show significant associations in Aim 1 will be tested for gene x environment interaction using the environmental measures described in the Appendix.

Aim 3: To test whether shared genetic or environmental factors can explain co-morbidity between dyslexia and SLI and whether it is possible to identify a correlation between genetic/environment background and specific sub-groups of phenotypes. The data generated in Aim 1 and Aim 2 will be used to assess whether it is possible to identify causative factors that either have a particular effect on specific phenotypes or affect different, broader areas of cognition. Factors that are found to have a pleiotropic effect across different phenotypes will be further explored to investigate a possible role in determining the observed co-morbidity between SLI, dyslexia and related disorders such as ADHD. Since both dyslexia and SLI could be hypothesised to be caused by an underlying phonological deficit it will be interesting to see if specific genes contribute to the variation of phonological outcomes. The identification of association between single factors and multiple phenotypes will lead to the question of whether it is possible to identify additional interacting factors that result in more distinctive phenotype. The ultimate aim of this analysis is to test whether it is possible to distinguish between well-defined cognitive deficits on the basis of determined genetic/environmental background.