Improved Hybrid de novo Genome Assembly, Gene Prediction and Annotation of Carrot (Daucus carota)

The release of the carrot genome sequence (v2.0) in 2016 has rapidly enhanced molecular and genomic research for this species, changing the nature of research in carrot biology. Research is shifting more towards extensive genetic screening for genome-wide association analysis and functional genomics. However, despite the high-quality of the current carrot genome assembly v2.0 release, improvements are needed due to the multiple challenges associated with short read sequencing data used to develop it. Here, an improved genome assembly, gene prediction and annotation of carrot DH1 (v3.0) is presented. The new genome assembly sequence covers 440 Mb, with a contig N50 >6Mb, all assembled into 9 pseudomolecules/chromosomes and 2 organellar genomes. Compared with the previous assembly (v2.0), the v3.0 assembly includes about 11% (54 Mb) of novel nucleotide sequence, >21% (>100 Mb) new sequences anchored at the chromosome level, and represents a >193 fold increase in contig N50. Using a combination of IsoSeq and Illumina transcriptome data, 36,216 gene models were predicted, with >4,000 additional gene models as compared to the previous gene prediction. Taking advantage of the IsoSeq full-length high-quality transcripts, >6,000 mis-predicted and partial gene models in v2.0 were identified and manually curated. Finally, a comprehensive catalog of alternative splicing (AS) events in the carrot DH1 was obtained, and efforts to identify tissue specific isoforms and AS events are ongoing.