ARS | Publication request: MARKER-ASSISTED SELECTION FOR MULTIPLE TRAITS IN CUCUMBER (CUCUMIS SATIVUS L.)

Submitted to: Meeting Proceedings
Publication Type: Proceedings
Publication Acceptance Date: August 20, 2003
Publication Date: October 15, 2003
Citation: Staub, J.E., Robbins, M.D., Fazio, G. 2003. Marker-assisted selection for multiple traits in cucumber (cucumis sativus l.). Meeting Proceedings.

Technical Abstract: The theoretical basis of marker-assisted selection (MAS), its potential advantages, and its integration with conventional breeding systems have been considered. However, practical application of MAS can only be justified when predicted benefits (long- and short-term gain from selection) outweigh the additional cost of MAS above traditional breeding methodologies. Four important cucumber yield components are earliness (days to anthesis; DA), sex expression (SE), multiple lateral branching (MLB), and fruit length [measured as length:diameter (L:D) ratio]. All of these characters are under genetic control from relatively few QTL (2-6 for each trait). These and other important yield components have been placed on a genetic map. We used this information for the design of a project which would test the applicability of multi-trait selection using MAS. A recurrent selection strategy was devised given the genetic information available on yield components (i.e., maps and gene action), the relatively tight linkage between target traits and marker loci, and the known correlation between traits and their heritabilities. Parental lines were chosen based on unique plant habits and fruiting characteristics, and included USDA lines 6632E, 6823B, 6947A, and 6995C. Integral to this strategy, four populations (1-4) were initially produced (F1) using bulk pollen from three lines crossed to a fourth line (e.g., 1 x 2,3,4; 2 x 1,3,4; etc.). Phenotypic selection, MAS, and a non-selected random control is being applied to each of these four populations. Gain from selection for the four traits over three cycles (C1-C3) will be estimated by the evaluation of populations in a multi-location replicated trial.