Molecular breeding methods
A Bayesian multilevel whole-genome regression model is disclosed and its prediction performance compared to that of the popular BayesA model applied to each population separately (no pooling) and to the joined data set (complete pooling). For small population sizes (e.g., <50), partial pooling increased prediction accuracy over no or complete pooling for populations represented in the estimation set. Partial pooling with multilevel models can make optimal use of information in multi-population estimation sets.
1. A method for selecting plants in a plant breeding program and for breeding selected plants, said method comprising:
a. constructing an optimized estimation data set for whole genome prediction and respectively selection by:
i. selecting and pooling predefined candidate sets of populations, said predefined candidate sets being selected from breeding crosses, and selecting plants within each breeding cross for phenotyping from a comprehensive list of plants targeted for selection wherein genotypic information is available for all candidate plants;
ii. computing, for a population of breeding crosses and breeding traits under active artificial selection, specific genetic similarity or divergence measures using objective functions of population and trait specific estimated QTL effects;
iii. constructing optimized estimation datasets in terms of number of populations pooled and respective plants within each population; and
iv. using partial pooling statistical models as a function of objective genetic similarity or divergence criteria;
b. phenotyping candidate plants in the optimized estimation data set;
c. genotyping breeding plants at a plurality of markers;
d. obtaining genomic estimated breeding values for the genotyped breeding plants using the phenotypes of the candidate plants in the optimized estimation data sets;
e. making selections within the complete set of genotyped candidate plants for selection plants based on the genomic estimated breeding values generated conditional on optimized estimation sets through partial pooling; and
f. breeding at least one of the selection plants with at least one other of the selection plants.
2. The method of claim 1 , wherein said genotypic information for the candidate plant is obtained via genotyping using SNP markers.
3. The method of claim 1 , wherein said breeding plants are homozygous.
4. The method of claim 1 , wherein the plant is selected from the group consisting of: maize, soybean, sunflower, sorghum, canola, wheat, alfalfa, cotton, rice, barley, millet, sugar cane and switchgrass.
5. The method of claim 1 , wherein the population is a genetically diverse population that includes individuals carrying one or more transgenes.
6. The method of claim 1 , wherein the population is a genetically diverse population that includes individuals with DNA edited with Cas9.
7. The method of claim 1 , wherein said genotypic information for the candidate is obtained by analyses of gene expression, metabolite concentration, or protein concentration.