Prediction of selection gains in Coffea canephora based on factorial scores

Abstract

The technique of factor analysis in the simultaneous selection of traits and prediction of genetic gains was evaluated in Coffea canephora var. conilon. Fourteen traits in 40 assessed genotypes were evaluated at two sites. The technique was used aiming at the structuring and simplification of the data, without information loss and with biological interpretation. The experimental design was of randomized blocks in four replications, each plot containing two useful plants. The technique was efficient for the data simplification and structuring. Moreover, the estimates of the predicted gains in the traits involved in the factors showed magnitude near the direct selection gain, attesting the suitability of the technique and its use in improvement programs of the species.