In this study real data from Iberian and Duroc individuals were used, as well as simulated F1, Iberian and Duroc backcrossed individuals. The ability of three clustering methods to separate the five populations was evaluated. Results indicate that the clustering methods do not detect the optimal partition. The problem is, in part, due to the different genetic configurations of the different strains of Iberian pigs that the methods can detect and, therefore, interfering in the separation between pure and crossed individuals.