¹Departments of Psychiatry and Pharmacology, Denver VA Medical Center and University of Colorado Health Sciences Center, 4200 East Ninth Avenue C-268-71, Denver, CO 80262, USA

Correspondence: R Freedman, E-mail: Robert.Freedman@UCHSC.edu

SIR – Meyer et al¹ conclude that CHRNA7, the gene that codes for the 7-nicotinic acetylcholine receptor, is excluded as a possible candidate gene in their extended pedigree of periodic catatonia, a subtype of schizophrenia. A recombination in one branch of their pedigree appears to exclude the region centromeric of D15S144 that includes CHRNA7. Mapping of disease loci based on a single recombination is a classic technique that is highly informative for many Mendelian traits, in which the assumption is that only a single rare allele accounts for the inheritance of the illness, but its value for complex, multigenic traits is less certain. Schizophrenia is assuredly such a multigenic trait;² even the periodic catatonia subtype is likely to be genetically heterogeneous, based on Meyer et al's finding that the linkage at 15q13–22 is largely accounted for by only a single family, with the remainder of their families having little evidence for linkage at any locus. The recombinant individual is one of two unaffected parents of a branch of the pedigree, so that the actual source of entry of the disease into this branch is uncertain. Indeed, all the three affected siblings in the branch share the same 15q13–22 haplotype from the other parent who has married into the extended pedigree, who is therefore also a possible source of the disease mutation.

A more informative pedigree in the NIMH Schizophrenia Genetics Initiative has a recombination in a schizophrenic proband that includes CHRNA7, and excludes D15S144 and more telomeric markers, the region that appeared to be included in the Meyer et al pedigree (Figure 1). In the NIMH pedigree, the half siblings are descended from an affected parent, and there is no other shared haplotype. The NIMH Genetics Initiative ascertainment rules did not allow pedigree extension through unaffected family members, to obviate the ambiguity found in the Meyer et al pedigree. D15S165, the most polymorphic marker near CHRNA7, shows significant transmission disequilibrium with schizophrenia in the NIMH families (P=0.0069) and allele 6 is specifically transmitted in African-American families (P=0.0077), such as the one shown in Figure 1.³

Figure 1.

Recombination in a pedigree from the NIMH Schizophrenia Genetics Initiative. Affected status is indicated by darkened symbols; subject numbers are shown below each affected individual with genotypes. The two unaffected parents were not available for genotyping. Subjects 001, 002, 003 met DSMIIIR criteria for schizoaffective disorder, depressed type; subject 004 met criteria for schizophrenia. CHRNA7 was genotyped by alleles of D15S1360, a dinucleotide repeat in the gene's second intron; although Meyer et al show both CHRNA7 and CHRNA7dup telomeric of D15S165, physical mapping in BAC and YAC contigs place CHRNA7dup centromeric of D15S165 and CHRNA7 telomeric.^4,5 Diagnostic and genotyping procedures were previously described.³

Full figure and legend (21K)

It is certainly possible that the Meyer et al and the NIMH pedigrees transmit schizophrenia by different genes in the 15q13–22 region or have cryptic recombinations that obscure genetic mapping or that either or both pedigrees have multiple entry of disease alleles from this or other loci. Thus, it would seem premature in a common, multigenic illness to claim exclusion of any locus based on a single recombination, particularly one that occurs in an unaffected individual.