RLE-illness contains four questions about serious illness, injury or problems occurring between the subject and close relatives, neighbours or the police. In the combined sample using additive (P=0.0156) and dominant (P=0.0088) model of heritability significant interactions with 5-HTTLPR were found, and these were not false discoveries according to the QVALUE programme (Table 2, Figure 1a). Independently in the Budapest sample using a dominant (P=0.0365) model we found a significant interaction; however, it did not survive the test for FDRs (Table 3) indicating no replication of our result in the two population subsamples.