We combine Sanger sequencing and ultra-high throughput sequencing to develop an annotated EST/gene sequence database for carrot. This sequence serves as a reference to identify Single Nucleotide Polymorphism (SNP) or insertion/deletion (Indel) variation in breeding germplasm of carrot using Illumina technology. The EST database serves as a reference for other Apiaceae species. The sequences and markers developed will form the basis for the first extensive gene-based genetic maps in this family.
| Name | Description |
|---|---|
A de novo assembly of carrot ESTs from four genetic backgrounds using the Illumina platform and Sanger sequencing. Normalized cDNA libraries were constructed from four sources: two orange unrelated inbred lines of European origin, B493 and B6274 with Imperator and Nantes root shapes, respectively; a purple/yellow inbred line B7262 derived from an intercross between purple Turkish and orange Danvers (European) carrots; and a pool of F₄ RILs derived from a cross between B493 and QAL, a wild carrot from North America. These pooled F₄'s are referred to as B493 × QAL and were derived from a single B493 × QAL F₁ plant. Therefore at most two haplotypes are represented among these transcripts. A total of 18,044 high quality Sanger sequences were obtained from the B493 library to generate a total of 8,221,411 nt with an average length of 456 nt. These sequences are deposited in GenBank under accession numbers JG753039 through JG771082. The three other libraries, B6274, B7262 and B493 × QAL were sequenced with Illumina GAII platform with 61 cycles to yield from 34 M to 39 M usable reads of 41 nt or longer for each genotype. These reads are available in the NCBI SRA archive under accession number SRA035376. A CAP3 assembly of B493 Sanger sequences produced 4,044 contigs plus 3,241 singletons (Additional file 1, Table S1). A multiple step assembly strategy was used to produce a de novo assembly of the three Illumina sequence sets. For each genotype two separate assemblies were produced using either Velvet combined with CAP3, or ABySS. The Velvet+CAP3 assembly gave 31,337, 34,218, and 39,901 contigs for B6274, B7262, and B493 × QAL, respectively. The number of contigs produced by ABySS assembly was higher, ranging from 133,933 in B6274 to 193 844 for B493 × QAL. To combine the four sequences sources (B493, B493 × QAL, B6272 and B7262), a combined CAP3 assembly was created of contigs ≥100 nt. This cut off was selected based on annotation frequency vs. contig length. The resulting sequence assembly produced 57,840 contigs plus 911 Sanger singletons with a total sequence length of about 45 Mb. The average length of the contigs and singletons was 768.2 nt and the N50 (the contig lengths for which 50% of the sequence in an assembly is in contig of this size or larger) was 1378 nt. Out of the 58,751 contigs and singletons, 6,912 (11.7%) contained B493 Sanger sequences. Among the Illumina-sequenced genotypes, B7262 sequences were most common in contigs, represented in 50,057 contigs (85.2%). Comparing Illumina-sequenced transcriptomes, a total of 19,762 contigs (38.1%) contained reads from only two genotypes, with 18.3% of the contigs having reads from B493 × QAL and B7262, 9.4% from B493 × QAL and B6274, and 10.4% from B7262 and B6274. More than 50% of the assembled contigs contained sequences from all three genotypes. B7262 had the highest number of genotype specific contigs (1,494, 2.9%), and B6274 had the lowest with 1,017 (2.0%) genotype specific genes. The sequences available for download are these 58,751 sequences from "Additional file 2" from the publication. |
| NCBI BioProject | |
|---|