Transcriptomic prediction of breeding values in loblolly pine.

Area of Science:

• Forestry science
• Quantitative genetics
• Molecular biology

Background:

• Phenotypic variation in forest trees is influenced by genetic and environmental factors.
• Heritability quantifies the genetic component of phenotypic variation.
• Genetic covariation modeling uses relationship matrices (pedigree or molecular markers) to predict genetic values.

Purpose of the Study:

• To test if shared gene expression or sequence variation patterns reflect genetic covariation among individuals with similar genetic values.
• To assess the predictive power of gene expression and SNP data for genetic covariance modeling.
• To determine if pedigree information is necessary for this modeling approach.

Main Methods:

• Collected gene expression data via high-throughput RNA sequencing from pooled seedlings of known parent genetic value.
• Analyzed data using various approaches, including selection of specific transcripts and SNP variation.
• Developed models incorporating transcript levels and SNP data.

Main Results:

• Selecting specific sets of transcripts improved model predictive power compared to using all transcripts or SNPs.
• Models combining transcript levels and SNP variation demonstrated higher predictive accuracy than models using only one data type.
• Pedigree information was not required for successful genetic variation modeling.

Conclusions:

• Shared patterns of gene expression and SNP variation can effectively model genetic covariation in forest trees.
• This method offers a powerful alternative to traditional genetic covariance modeling, especially for natural populations.
• The findings have significant implications for tree breeding and conservation genetics.