The norepinephrine transporter (NET) gene is an attractive candidate gene for attention-deficit hyperactivity disorder (ADHD). Noradrenergic systems are critical to higher brain functions such as attention and executive function, which are defective in ADHD. The clinical efficacy of medications that target NET also supports its role in the etiology of ADHD. Here, we have applied a dense mapping strategy to capture all genetic variations within the NET gene in a large number of ADHD families (474 trios). As a result, we found association of the same alleles from two single-nucleotide polymorphisms (rs3785143 and rs11568324) previously identified in another large-scale ADHD genetic study (International Multisite ADHD Geneproject). Furthermore, the effect sizes were consistent across both studies. This is the first time that identical alleles of NET from different studies were implicated, and thus our report provides further evidence that the NET gene is involved in the etiology of ADHD.

Although the concept of meta-analysis of multiple linkage scans of a genetic trait is not new, it can be difficult to apply to published data given the lack of consistency in the presentation of linkage results. In complex inheritance common diseases, there are many instances where one or two studies meet genome-wide criteria for significant or suggestive linkage but several other studies do not show even nominally significant results with the same region. One possibility for resolving differences between study results would be to combine an available result parameter of several studies. We describe here a method of regional meta-analysis, the multiple-scan probability (MSP), which can be used on published results. It combines the reported P-values of individual studies, after correcting each value for the size of the region containing a minimum P-value. Analyses of the power of MSP and of its type I error rates are presented. The type I error rate is at least as low as that for a single genome scan and thus genome-wide significance criteria may be applied. We also demonstrate appropriate criteria for this type of meta-analysis when the most significant study is included, and when that study is used to define a region of interest and then excluded. In our simulations, meta-analysis is at least as powerful as pooling data. Finally, we apply this method of meta-analysis to the evidence for linkage of autism susceptibility loci and demonstrate evidence for a susceptibility locus at 7q.