The primary activities in a cotton breeding program involve germplasm storage, identification, parental selection, evaluation of segregating populations in breeding nurseries, evaluation of inbred progenies in replicated trials, collection and summarization of data, and printing of fieldbooks, labels, and maps. Many or all of these activities are handled by data base management computer programs adapted to personal computers that are commonplace in many research programs. AGROBASE/4(TM) (1) is a comprehensive program for data management and data analysis that can be implemented in a cotton improvement program. The package was developed in dBASE(TM)-compatible FOXPRO(TM).

The AGROBASE commands are grouped into five main catogories: file management, data management, experimental designs and statistical analysis, and plant breeding. The desirable feature of AGROBASE is the potential to execute all necessary functions of a breeding program with a single software package.

The ACCESN command is used to create a "master germplasm" file that facilitates germplasm and pedigree data management. This master file contains all near-homozygous lines or cultivars existing in the breeding program. The HYBRID command then is used to create segregating populations that enter the breeding program. This command utilizes germplasm from the master file and their respective pedigree information. SEGPOP command is used to generate F(n+1) files from segregating F(n) files according to the pedigree breeding method employed. All pedigree, agronomic, and fiber quality data from each progeny selected is maintained in the new F(n+1) file. All nursery fieldbooks and labels are printed to contain the necessary pedigree and previous generation data. Standard or custom book printing functions are available. In the F(5) or F(6) generation, progeny are bulk-harvested for entry into preliminary replicated trials. DESIGN command extracts selected entries, plus pedigrees and 3 or 4 generations of data, from the F(n) files and creates replicated trials of the appropriate experimental design for statistical analysis. All selected bulk progenies are also entered into the "master germplasm" file. Statistical analysis of the replicated trials are conducted by standard ANOVA of RCBD or LATTICE designs. Nearest neighbor analysis (2) using adjacent residuals is employed to adjust for spatial trends in the data. COLLECT command is finally utilized to summarize data over years and experiments for reports and germplasm release.

References

1) Mulitze, D.K. 1990. AGROBASE/4: A microcomputer database management and analysis system for plant breeding and agronomy. Agron. J. 82:1016-1021.

2) Stroup, W.W., Baenziger, P.S., and D.K. Mulitze. 1994. Removing spatial variation from wheat yield trials: a comparison of methods. Crop Sci. 34(1):in press.