Estimation of sampling variance of molecular marker data using the bootstrap procedure

Area of Science:

• Plant genetics and breeding
• Molecular markers
• Bioinformatics

Background:

• Understanding genetic relationships among genotypes is crucial for efficient plant germplasm organization and sampling in breeding programs.
• Genetic distance (GD) estimation using random restriction fragment length polymorphisms (RFLPs) is a key tool for assessing these relationships.
• Accurate GD estimation requires knowledge of the sampling variance associated with RFLP markers to determine the necessary number of markers for desired precision.

Purpose of the Study:

• To estimate the sampling variance for GD among 37 maize inbred lines using 1202 RFLPs.
• To determine the required sample size of RFLPs for a given level of precision in GD estimation.
• To analyze the relationship between the coefficient of variation (CV) for GD and the number of RFLP markers used.

Main Methods:

• Estimated sampling variance for GD using 1202 RFLPs generated from 251 enzyme-probe combinations (EPCs) across 37 maize inbred lines.
• Utilized bootstrap sampling to assess the impact of sample size on the coefficient of variation (CV) of GD.
• Compared the rate of CV reduction using individual polymorphic bands versus EPCs as sampling units, and analyzed the effect of different restriction enzymes (BamHI, EcoRI, HindIII).

Main Results:

• The coefficient of variation (CV) for GD showed a nearly linear relationship between its expected standard deviation and mean.
• The decrease in mean CV with increasing numbers of RFLP bands was dependent on the sampling unit (bands vs. EPCs) and the relatedness of the inbred lines.
• The rate of CV reduction with increasing sample size was consistent across restriction enzymes when using individual polymorphic bands, but intercepts differed when using EPCs due to varying band numbers per EPC.

Conclusions:

• The study provides a framework for determining the optimal number of RFLP markers needed for precise genetic distance estimation in maize breeding.
• Sampling units (individual bands vs. EPCs) and enzyme choice significantly influence the precision and efficiency of GD estimation.
• Understanding these sampling variances allows for more informed decisions in germplasm organization and marker selection for plant breeding programs.