Background: The study of gene-environment interactions (G by E) is one of the most promising strategies to uncover the origins of mental disorders. Replication of initial findings, however, is essential because there is a strong possibility of publication bias in the literature. In addition, there is a scarcity of research on the topic originated from low- and middle-income countries (LMIC). The aim of this study was to replicate G by E hypotheses for externalizing problems among adolescents in a middle-income country. Methods: As part of the Pelotas 1993 Birth Cohort Study, 5,249 children were enrolled at birth and followed up to the age of 15 years, with an 85.7% retention rate. We sought an interaction between the homozygosity of the 10-repeat allele at the dopamine transporter ("DAT1") gene and prenatal maternal smoking in the development of hyperactivity problems during adolescence assessed by the Strengths and Difficulties Questionnaire. We also tested for an interaction between the uVNTR polymorphism at the monoamine oxidase A ("MAOA") and the experience of childhood maltreatment in the occurrence of conduct problems among adolescent boys. Results: Although there was a clear association between prenatal maternal smoking and hyperactivity scores in adolescence ("p" less than 0.001), no main genetic or interaction effects for the "DAT1" gene were detected. Similarly, childhood maltreatment showed to be associated with conduct problems among boys ("p" less than 0.001), with no observable main genetic or interaction effects for the "MAOA" gene. Conclusions: In the largest mental health G by E study performed in a LMIC to date, we did not replicate previous positive findings from the literature. Despite the presence of main environmental effects, there was no evidence of effect modification by genotype status. Additional replication efforts to measure G by E are needed to better understand the origins of mental health and illness, especially in LMIC. (Contains 1 table and 2 figures.)