A shift towards larger CAG/CTG triplet repeats and schizophrenia (SCZ) and bipolar affective disorder (BPAD) has been detected by several recent studies, using the Repeat Expansion Detection (RED) technique, however no specific loci have been shown to be responsible for this shift. Further analyses by our group of RED (CTG)10 ligation products amongst an extended sample of patients and comparison with controls matched for age, sex and ethnicity show no significant differences in distribution (P= 0.23, n=95; P=0.93, n=91, for SCZ and BPAD respectively). Alleles at two recently discovered unstable trinucleotide repeat loci at 18q21.1 (SEF2-1B) and 17q21.3 (ERDA1) have also been analysed in affecteds and matched controls. We observed no increase in frequency of larger alleles (>37 repeats) in affected individuals at SEF2-1B (BPAD: P=0.95, n= 100; SCZ: P=0.61, n=97) or at ERDA1 (BPAD: P= 0.4, n = 101; SCZ: P= 0.05, n = 151, with larger alleles more frequent in controls). Our findings suggest that larger CAG/CTG repeats at these loci are neither major contributory factors to the etiology of psychosis, nor in linkage disequilibrium with a gene that is. Furthermore, when the RED results were compared to allele sizes at SEF2-1B and ERDA1, it was observed that a majority of SCZ, BPAD and control individuals with large RED products had a large allele at either or both sites (78% for RED products > or =270 bp; 62% for RED products > or =180 bp).