The Complete Apple Genome: Evolutionary Events Revealed

Accurate sequence information and genome assemblies are critical for studies on genomic and epigenomic variations. Although partial genome information is already available for apple (Malus domestica Borkh), the assembled sequence is fragmented and lacks repeated regions. Using second- (Illumina) and third-generation sequencing and optical mapping technologies (PacBio and BioNano), we have generated a high-quality genome assembly of a 'Golden Delicious' doubled haploid tree. Our de novo assembly resulted in a genome of 649.7 Mb, with a N50 of 5.6 Mb. Seventeen pseudo-chromosomes were constructed, representing 96.3% of the sequence data, and validated using a high density integrated genetic linkage map and linkage disequilibrium analysis. Using mRNA and EST sequence data, gene prediction modelling identified 42,140 protein coding genes. Repeat sequences, representing 57.3% of the assembly, were annotated and provide an unprecedented opportunity to investigate the 'dark matter' of the genome. We have identified a new, hyper-repetitive LTR-retrotransposon sequence named HODOR which is over-represented in heterochromatic (presumably centromeric) regions. Furthermore, we present the results of a differential analysis of DNA methylation patterns between apple leaves and fruits, and show differentially methylated regions located close to promoters of genes that are potentially involved in fruit development. Finally, we show that a major burst of transposable elements occurred around 21 My ago, which might have contributed to the divergence between apple and pear. The greatly improved apple genome assembly presented here provides important insights into the apple genome structure and thereby lays the foundation for future genomic and epigenomic studies in Rosaceae.