Bridging The Gap Between Sequence and Biology Of Apple

Genome annotation is a complex endeavor despite advances in the quality of gene prediction software. In an effort to develop high-quality gene model for apple (Malus x domestica), we aligned massively parallel cDNA sequencing reads of 19 apple tissues to the v.1 apple genome. Integration of RNA-Seq allowed us to identify hundreds of transcripts missed by the automatic annotation procedure, to correct predicted gene structures, and to detect a multitude of alternative and trans-splicing events. Furthermore, data from the v.1 apple genome allowed us to develop a NimbleGen microarray chip with probes complementary to both sense and anti-sense transcripts. Anti-sense RNA expression has recently been shown to be a common regulatory mechanism for gene expression. Our expression analysis over eight apple organs show that anti-sense RNAs were identified in up to 60% of expressed genes. Interestingly, the percentage of anti-sense RNA identified varied among gene and transcription factor families. In plants, the long dsRNA formed by hybridization of sense and anti-sense transcripts are cleaved into double stranded siRNAs by Dicer proteins. To further validate anti-sense transcripts expression, we sequenced and aligned smallRNAs (siRNA and miRNA) to apple CDS and compared anti-sense RNA expression values and smallRNA reads coverage for a particular library.

As shown in this study, sequencing projects require continual improvements and refinements. Our RNA-seq approach combined with large-scale transcript expression analysis by microarray provided a powerful method to rapidly improve the apple genome annotation and build a comprehensive expression atlas of both sense and anti-sense transcripts.