Improved Hybrid de novo Genome Assembly, Gene Prediction and Annotation of Carrot

This analysis represents the JBrowse annotation of gaps in the assembled carrot genome sequence "Carrot Genome Assembly DH1 v3.0".

This JBrowse track shows all gaps in the genome assembly consisting of runs of 100 'N's, which represent gaps of unknown size. There are 554 gaps in the entire genome assembly.

PlantTFcat, a reference plant transcription factor and transcriptional regulator categorization tool, was used to predict the transcription factors and regulatory genes in v3 gene models as well as the DCARv2 genes for comparison purposes. The data from this analysis is presented in supplementary table S17 in the carrot genome publication.

PRGdb 3.0, a comprehensive platform for prediction and analysis of plant disease resistance genes, was used to predict the disease resistant genes in v3 as well as the v2 for the comparison purposes. The data from this analysis is presented in supplementary table S19 in the carrot genome publication.

SSRs were detected in the Daucus carota V.3 assembly using misa, and when possible primers were designed with primer3.

misa detection parameters 2-6 3-4 4-3 5-3 6-3 7-3 8-3 and interruptions 10

This analysis has not been published.

RESULTS OF MICROSATELLITE SEARCH
================================

Total number of sequences examined:                  11
Total size of examined sequences (bp):      441,119,442
Total number of identified SSRs:                148,264
Number of SSR containing sequences:                  11
Number of sequences containing more than 1 SSR:      11
Number of SSRs present in compound formation:    12,941


Distribution to different repeat type classes
---------------------------------------------

Unit size  Number of SSRs
2          52,385
3          30,622
4          38,281
5          16,265
6           7,217
7           2,622
8             872

128,207 designed primer pairs

A multi-step approach was used to predict the most comprehensive gene model catalog for the carrot genome v3. MAKER and GeMoMa were used to perform gene prediction based on the integration of de novo gene prediction and evidence-based predictions. For MAKER, carrot ESTs, DH1 Illumina and IsoSeq transcriptome sequences, gene models obtained from five closely related or model species and proteins from Uniprot-sprot were used as transcript evidence. AUGUSTUS v2.5.5 and SNAP were used for de novo prediction. Through this analysis MAKER predicted 28,721 gene models. Next, GeMoMa was used to improve the quality of the splice junction sites predicted by MAKER and to predict the gene models that were not predicted by MAKER. The datasets included as input in GeMoMa were: predicted genes from the five related species or model species used for MAKER prediction; final gene models produced from MAKER pipeline; splice sites mined from the mapping of the DH1 Illumina transcriptome data on DH1 v3. This analysis produced an intermediate set of 32,625 gene models. A final step was performed to refine all gene models and predict any missing models. In this step, gene models predicted on the DH1 v2 assembly, named DCARv2 (32,112) and RefSeq (44,484) were transferred/re-predicted to the DH1 v3 genome assembly using GMAP and GenomeThreader. DCARv2 or RefSeq gene models that were not predicted by MAKER+GeMoMa, that had experimental evidence an that were not masked, were considered as new gene models. In those cases where the structure of the RefSeq and DCARv2 gene models were not in agreement, the correct structure was manually inspected using the experimental evidences. Finally, high-quality IsoSeq transcripts were mapped to the DH1 v3 assembly using GMAP and GenomeThreader, and those transcripts mapping with appropriate gene structure and not predicted in the previous steps, were added to the gene model catalog.