Explain about genomic selection populations of different types?

Genomic selection (GS) utilizes genomic information to predict the breeding value of individuals for selection purposes. Different types of populations are used in genomic selection, each with its own characteristics and applications.

Training Population: This population is used to calibrate the prediction model in genomic selection. It typically consists of a large number of individuals with both phenotypic and genotypic data. The genomic prediction model is trained using this population to establish the relationship between markers and phenotypes, allowing for accurate prediction of breeding values in subsequent generations.

Validation Population: The validation population is used to evaluate the accuracy of genomic predictions generated by the model trained on the training population. It is separate from the training population and often represents individuals from different genetic backgrounds or environments. By comparing predicted breeding values with observed phenotypes in the validation population, the reliability of the genomic prediction model can be assessed.

Breeding Population: The breeding population comprises individuals selected for further breeding based on their predicted breeding values generated through genomic selection. These individuals are chosen to serve as parents for the next generation, aiming to improve target traits such as yield, disease resistance, or quality. Genomic selection allows breeders to make more informed decisions about which individuals to select as parents, potentially accelerating genetic gain in breeding programs.

Reference Population: The reference population is a subset of the training population that serves as a reference for imputing missing marker data in individuals with incomplete genotypic information. Imputation is a process of inferring missing genotypes based on known marker-trait associations in the reference population. A large and diverse reference population enhances the accuracy of imputation and subsequently improves the performance of genomic selection models.

Bi-parental Population: Bi-parental populations are created by crossing two parental lines with contrasting traits of interest. These populations are often used for genetic mapping, QTL (Quantitative Trait Locus) identification, and marker discovery. In genomic selection, bi-parental populations can be used to develop training populations for specific breeding objectives or to validate genomic prediction models in controlled environments.

Each type of population plays a distinct role in the genomic selection process, contributing to the development of accurate prediction models and the improvement of target traits in breeding programs. By utilizing these populations effectively, breeders can harness the power of genomics to accelerate genetic gains and enhance crop or livestock improvement.