The sample consists of individuals who previously participated in the ‘CoLaus’ study 
a community survey of 6 738 randomly-selected residents of the city of Lausanne (Switzerland). Of the 3 400 Caucasians who accepted the psychiatric evaluation (55% participation rate), 2 239 individuals also completed and returned the EPQ questionnaire (66% return rate), and had genotype data available.
The mean age of the study individuals was 51.9 (SD 8.85; range 36–70), and 55.4% were female. These compare with mean age 51.2 (SD 8.82; range 36–70) and 53.0% female among all 3,307 PsyCoLaus individuals, indicating that responders to the questionnaire are slightly older and more likely to be female than non-responders.
A QQ plot of the p
-values for the genotyped SNPs (supplementary material File S1
) suggested negligible inflation due to population structure or cryptic relatedness (genomic control λ
1.014), but also little evidence of p
-values below those expected under the null hypothesis. Results from the main-effects analyses are shown in , and . Both genotyped and imputed SNPs within the gene ARRDC4
, at the distal end of 15q, almost reached p
. We identified several imputed SNPs with p
on 6q, within the gene NKAIN2
, which has not previously been reported in linkage or association studies, although the strongest association at a genotyped SNP (rs11154221) in this gene has p
. No SNP showed a notable interaction with sex (). SNP by age interaction analyses revealed multiple associations at the distal end of 13q within gene GPC6
(, ). The strongest signal (p≈10−7
) was at a genotyped SNP, rs9561329; EPQ score tends to increase with age for carriers of the minor allele, and decreases for non-carriers (). Conditional regression analyses revealed that the signal at each of the SNPs reported in effectively disappeared when the most strongly-associated SNP at that locus was included as a predictor in the regression, suggesting at most one causal variant underlying each signal.
Most significant main effect and interaction results for association with neuroticism (EPQ-N).
EPQ-N distribution by 3-year age groups according to minor allele carrier status at rs9561329.
SNP rs702543 in gene PDE4D
, reported by Shifman 
to be associated with neuroticism, was not present among our 1.7 M genotyped or imputed SNPs. However, we obtained p
0.006 for association at the imputed SNP rs296410 that showed the highest R2
with rs702543 in the Hap Map CEU samples (R2
0.51); in our study, the major allele (T, frequency 0.52) at rs296410 shows a positive linear association with EPQ-N score. We could not replicate the association found by van den Oord 
for the gene MAMDC1
on 14q. Three of the four SNPs they reported had a perfect proxy in our SNP set, but our lowest p
-value in this gene was p
0.43. We tested 3 genotyped SNPs in the gene NPY
studied by Zhou30
, including a SNP, rs16126, in high LD (R2
≈0.9) with two of their reported SNPs, but we found p
>0.4 at all three SNPs. Finally, we found no significant enrichment of a GO functional category among SNPs with p
A separate analysis was carried out on the X chromosome, including the pseudo-autosomal region. Genotypes were coded as 0, 1, and 2 for females, and 0 and 2 for males, and sex was included as a covariate in the linear regression model. No significant associations were found.