Figure 1. SLAF-seq flowchart.

i) Pre-design scheme for SLAF selection using training data. The reduced representation design must be decided based on marker efficiency characteristics, which include random distribution throughout the genome, uniqueness in the genome, and consistent amplification efficiency among selected markers. A pilot experiment was performed to evaluate the amplification efficiency based on the pre-designed scheme. ii) SLAF-seq library construction. Genomic DNA was digested by groups of enzymesdesigned for individuals. Double barcodes were added to two round PCR reactions to discriminate each individual and to facilitate the pooling of samples for size selection, which maintained consistent fragment size among individuals. iii) Deep sequencing for the pooled RRLs with the Illumina paired-end sequencing protocol, and genotype definition and validation by software.