Hyperpolarization-activated cyclic nucleotide-gated (HCN) channels are encoded by four genes (HCN1–4) and, through activation by cyclic AMP (cAMP), represent a point of convergence for several psychosis risk genes. On the basis of positive preliminary data, we sought to test whether genetic variation in HCN1–4 conferred risk of depression or cognitive impairment in the Generation Scotland: Scottish Family Health Study. HCN1, HCN2, HCN3, and HCN4 were genotyped for 43 haplotype-tagging SNPs and tested for association with DSM-IV depression, neuroticism, and a battery of cognitive tests assessing cognitive ability, memory, verbal fluency, and psychomotor performance. No association was found between any HCN channel gene SNP and risk of depression, neuroticism, or on any cognitive measure. The current study does not support a genetic role for HCN channels in conferring risk of depression or cognitive impairment in individuals from the Scottish population.

Bipolar disorder, schizophrenia and recurrent major depression are complex psychiatric illnesses with a substantial, yet unknown genetic component. Linkage of bipolar disorder and recurrent major depression with markers on chromosome 4p15–p16 has been identified in a large Scottish family and three smaller families. Analysis of haplotypes in the four chromosome 4p-linked families, identified two regions, each shared by three of the four families, which are also supported by a case-control association study. The candidate gene phosphatidylinositol 4-kinase type-II beta (PI4K2B) lies within one of these regions. PI4K2B is a strong functional candidate as it is a member of the phosphatidylinositol pathway, which is targeted by lithium for therapeutic effect in bipolar disorder. Two approaches were undertaken to test the PI4K2B candidate gene as a susceptibility factor for psychiatric illness. First, a case-control association study, using tagging SNPs from the PI4K2B genomic region, in bipolar disorder (n = 368), schizophrenia (n = 386) and controls (n = 458) showed association with a two-marker haplotype in schizophrenia but not bipolar disorder (rs10939038 and rs17408391, global P = 0.005, permuted global P = 0.039). Second, expression studies at the allele-specific mRNA and protein level using lymphoblastoid cell lines from members of the large Scottish family, which showed linkage to 4p15–p16 in bipolar disorder and recurrent major depression, showed no difference in expression differences between affected and non-affected family members. There is no evidence to suggest that PI4K2B is contributing to bipolar disorder in this family but a role for this gene in schizophrenia has not been excluded.

Background

Regions of interest identified through genetic linkage studies regularly exceed 30 centimorgans in size and can contain hundreds of genes. Traditionally this number is reduced by matching functional annotation to knowledge of the disease or phenotype in question. However, here we show that disease genes share patterns of sequence-based features that can provide a good basis for automatic prioritization of candidates by machine learning.

Results

We examined a variety of sequence-based features and found that for many of them there are significant differences between the sets of genes known to be involved in human hereditary disease and those not known to be involved in disease. We have created an automatic classifier called PROSPECTR based on those features using the alternating decision tree algorithm which ranks genes in the order of likelihood of involvement in disease. On average, PROSPECTR enriches lists for disease genes two-fold 77% of the time, five-fold 37% of the time and twenty-fold 11% of the time.

Conclusion

PROSPECTR is a simple and effective way to identify genes involved in Mendelian and oligogenic disorders. It performs markedly better than the single existing sequence-based classifier on novel data. PROSPECTR could save investigators looking at large regions of interest time and effort by prioritizing positional candidate genes for mutation detection and case-control association studies.

We have compared the accuracy, efficiency and robustness of three methods of genotyping single nucleotide polymorphisms on pooled DNAs. We conclude that (i) the frequencies of the two alleles in pools should be corrected with a factor for unequal allelic amplification, which should be estimated from the mean ratio of a set of heterozygotes (k); (ii) the repeatability of an assay is more important than pinpoint accuracy when estimating allele frequencies, and assays should therefore be optimised to increase the repeatability; and (iii) the size of a pool has a relatively small effect on the accuracy of allele frequency estimation. We therefore recommend that large pools are genotyped and replicated a minimum of four times. In addition, we describe statistical approaches to allow rigorous comparison of DNA pool results. Finally, we describe an extension to our ACeDB database that facilitates management and analysis of the data generated by association studies.

Once thought to be impossible or a waste of resources, the initial high-volume stages of sequencing the human genome have been completed.