High-Throughput Genotyping in Eucalyptus Based on Targeted Sequencing: Applications in Genomic Selection

Methods that require genome-wide analysis, such as genomic selection and association genetics, create a need for new approaches of high-throughput genotyping of large numbers of markers. With the continuous decrease in sequencing costs, one approach is to sequence large numbers of targeted regions distributed throughout the genome, and detect variants based on read alignment. Here we present a high-throughput genotyping approach for Eucalyptus based on sequence capture. We designed probes to target 40,000 evenly spaced loci in the E. grandis genome. These probes were used to capture these regions in 700 individuals belonging to a highly heterozygous breeding population made of interspecific hybrids. For each individual, barcoded libraries were prepared and the product of the capture reaction was sequenced using Illumina HiSeq platform. Sequences were aligned to the E. grandis reference genome and SNPs identified in the population. Using the described method, we generated a dataset with a median sequencing depth of 24x, and we were able to identify 43,681 polymorphic markers. In addition, one given sample was repeated 14 times, allowing us to estimate a genotypic call repeatability of 0.997. The marker data obtained for the population was used to generate genomic selection prediction models for tree diameter and height and wood specific gravity. The models were developed in a cross-validation design using random regression best linear unbiased predictor (BLUP) and resulted in accuracies ranging from 0.59 to 0.76. These accuracies indicate the feasibility of this genotyping approach to Eucalyptus breeding.