Whole Genome Amplification of Single Pollen Grains and SNP Array Data for Accurate Haplotype Determination in Citrus

Haplotypes are an essential feature of the genome and many findings have indicated that relationships between DNA sequence and phenotype can be better understood with haplotype information. However, for outbred organisms there are no cost-effective approaches to obtain phased genotypes for a large number of individuals and loci. To overcome the existing technological impediments to obtaining chromosome-length haplotype information in heterozygous individuals we developed an easy method for single pollen grain isolation, lysis and whole genome amplification (WGA) by optimizing the REPLI-g single cell kit (QIAGEN). To reveal the haplotypes of the donor plant, WGA DNA of pollen grains was processed on the Axiom™ Citrus56AX Array (Affymetrix, Inc.) which includes 57,933 autosomal and 500 chloroplast SNPs. 273 WGA of single pollen grains from 39 diploid accessions were genotyped and alleles called with the Axiom Analysis Suite. Genotype calls were provided if the confidence value for the calls were under a threshold of 0.15 (default); otherwise, the genotype was assigned "No Call". WGA Average Cluster Call Rate was equal to 99.06%. For haplotype estimation Phase 2.1, Bayesian statistical method to infer haplotype frequencies from diploid data, was used. In order to facilitate the use of this program on haploid data, haploid chromosome data were randomly paired to form pseudo homologous chromosomes. A total of 10 pseudo diploid individuals were analyzed for each accession. As an independent validation, the predicted phase of Fairchild mandarin was compared to the phase obtained by linkage mapping with JoinMap version 4.1. The phase predicted by PHASE analysis of pollen was the same as found by linkage mapping.