The importance of the genotype×environment interaction for selection and breeding programmes in tropical conditions.

The literature on genotype × environment interaction (I&#60;sub&#62;GE&#60;/sub&#62;) was reviewed and discussed in order to obtain an overview of its effects for breeding programmes in tropical conditions. The results from cattle were the example used, owing to the great number of references available. The evidence shows that the performance of a specialized dairy breed introduced in the tropics is very far from the expected results and the best sires in the exporter countries (developed countries) were not the best in the tropical countries (importer). This same trend was found when the results from <i>Bos taurus</i> × <i>Bos indicus</i> crossbreeding programmes were examined. The publications for beef cattle show ahigh level of genetic correlations between the breeding value of the sires estimated in different countries. However, when the level of connection between animals was taken into account, the effect of I&#60;sub&#62;GE&#60;/sub&#62; was evident. The importance of I&#60;sub&#62;GE&#60;/sub&#62; was also found between countries with a high level of concentrate versus pasture feeding base: the same results were published for tropical conditions. A very large dataset of milk production statistics for the tropics (more than 200 000 records) was analysed in order to present some results on I&#60;sub&#62;GE&#60;/sub&#62; based on a more robust statistical procedure. In this study, the milk production level of the herd was used as an ‘environmental variable’ and the genetic parameters were estimated by a classical simple sire model, a character state model (CSM) and a reaction norm model (RNM). The results from RNM were considered to be the best, and allowed the estimation of the evolution of genetic variance in the entire trajectory of environment (2200–5000 kg of milk production). A clear effect of I&#60;sub&#62;GE&#60;/sub&#62; was evident and the best sires in a low environmental level were not the best in a high environmental level. A large genetic variability in plasticity was detected. Accordingly, the use of RNM is highly recommended in tropical breeding programmes because it allows the identification of the more adapted sires to heterogeneous environment. More research is needed for a better understanding of I&#60;sub&#62;GE&#60;/sub&#62; in the tropics, particularly more collaboration between tropical countries is necessary ... it is a dream to establish a tropical Interbull that enables us to reach the same results as those obtained in developed countries!