Plant materials and propagation
The doubled haploid onion line ‘CUDH2150’ was provided by Cornell University [30, 31] and the heterozygous landrace ‘Nasik Red’ (PI271311) was obtained from the USDA ARS Plant Genetics Resources Unit (Cornell University, Geneva, NY). Two individual flowering plants were cross-pollinated by blowflies and multiple F1 plants were individually self-pollinated to generate F2 families. Two F1 plants spontaneously produced topset bulbils, which were replanted and mass-pollinated to provide two very large F2 families. Samples of these families were grown at Lat 42 deg S near Christchurch, New Zealand. Cured bulbs were phenotyped for red bulb colour and freeze-dried samples were analyzed for fructan and hexose content as described elsewhere . DNA was isolated from fresh leaf material or freeze-dried bulb tissue as described previously . Working sets of PCR templates were generated from master stocks by whole-genome amplification using GenomiPhi V2 (GE Healthcare).
Total RNA was extracted from leaves and shoot meristem at the 4–5 leaf stage, prior to commencement of bulbing, from multiple plants of ‘CUDH2150’ and ‘Nasik Red’. Poly-A RNA was purified using Ambion Poly (A) Purist Kit (Life Technologies), as per manufacturers’ protocol.
cDNA synthesis was performed using the MINT cDNA Synthesis Kit (Evrogen). First strand synthesis was carried out on 2 μg polyA+ RNA substituting the kit 3′ primer with the modified primer 5′AAGCAGTGGTATCAACGCAGAGT(5)GT(9)CT(10)VN 3′. Then ds cDNA synthesis was performed with the additional 3′ primer 5′AAGCAGTGGTATCAACGCAGAGT(5)GTCT(4)GTTCTGTTTCT(4)VN at equimolar concentration to the kit “PCR Primer M1”. The optimal number of cycles was determined at 19 for Onion cDNA and 24 cycles for the kit control. After cDNA synthesis, ds cDNA was purified using the High Pure PCR Product Purification Kit (Roche). Approximately 3 μg ds cDNA was recovered from onion and 1.6 μg from the kit control. Normalization of cDNA was carried out with the Trimmer cDNA Normalization Kit (Evrogen) using 1.3 μg ds cDNA. The optimal number of cycles for the first amplification of normalized cDNA, was determined at 10 and the second amplification was performed for a total of 12 cycles. Approximately 8 μg of normalized cDNA was synthesized for sequencing. GS-FLX standard libraries were prepared from each genotype using unsheared cDNA and each was sequenced on 1/16 of a plate. Normalisation was assessed by BLASTN/X comparisons with Onion Gene Index V2.0 , rice and Arabidopsis unigene sets. A GS-FLX Titanium library was synthesized from the ‘CUDH2150’ cDNA and sequenced on a full Titanium plate. The ‘Nasik Red’ GS-FLX standard library was sequenced on full GS-FLX plate. Sequence data are accessible at NCBI under BioProject 60277. Raw flowgram data was submitted to Genbank SRA (Accession SRX031644-6).
Bioinformatics and marker design
A reference assembly of ‘CUDH2150’ was generated by assembling adapter-trimmed reads (SRA SRX031645) using Roche Newbler V 2.0.01.14 with options -cdna -cpu 6 -minlen 45 -tr -rip -icl 100. Reads showing significant BLASTN homology (E < 10-10) to plant ribosomal RNA sequences were excluded from the assembly. Contigs from the assembly were filtered by length and quality using Prinseq  to meet the Genbank Transcriptome Shotgun Assembly (TSA) standards and submitted to TSA as accessions JR842819 – JR863573.
Polymorphisms were detected by mapping ‘Nasik Red’ reads onto the ‘CUDH2150’ reference assembly using Roche gsMapper with default parameters. Tools for parsing gsMapper 454HCDiffs.txt/454AllDiffs.txt variant output files, detecting restriction polymorphisms and performing bulk PCR primer design were developed using GNU awk, Perl and BioPython  and then adapted for use in the Galaxy bioinformatics framework [16–18]. These scripts along with additional helper scripts for primer analyses and format conversions are freely available for download at Github (https://github.com/cfljam/galaxy-pcr-markers/) and for installation into Galaxy at the Galaxy Toolshed (http://toolshed.g2.bx.psu.edu) as repository ‘pcr_markers’.(http://toolshed.g2.bx.psu.edu/repos/john-mccallum/pcr_markers/). Amplicon size of 90–120 bp was used for design of CAPS markers, and 60–100bp for indel and HRM markers. HRM design was limited to class I and II SNPs  through filtering with standard Galaxy tools.
Initial screens of the SNP and indel markers were carried out using templates from ‘Nasik Red’, ‘CUDH2150’ and the F1 parent of the F2 population. Markers that were heterozygous in the F1 and segregating in an F2 subset of 9 lines were then tested on a core set of 93 F2 lines. Markers were assessed as multi-locus if multiple fragments were present after amplification with ‘CUDH2150’.
Markers were amplified by PCR using 0.5 U ThermoPrime Taq DNA polymerase (Thermo Fisher Scientific) in 15 μl reactions containing 1x PCR buffer, 200 μM dNTP, 1.5 mM MgCl2, 0.5 μM each primer and 20 ng template DNA. Amplifications carried out on a MasterCycler epGradientS (Eppendorf). The conditions included an initial denature at 95°C for 2 min then 40 cycles of 95°C for 30 s, 55°C for 30 s and 72°C for 30 s with a final extension of 7 min at 72°C. For CAPS markers the PCR products (5 μl) were digested in a 10 μl reaction using 3 U of restriction enzyme (NEB) (Additional file 1: Table S1) with the appropriate buffer at 1X final concentration and BSA where necessary. The digests were incubated for 3 h at 37°C or 65°C for TaqI digests. PCR and digestion products were separated using electrophoresis with a 4% agarose gel (2% Seakem LE + 2% NuSieve 3:1) and visualised under UV after ethidium bromide staining.
HRM markers were amplified in a 10 μl reaction using 1x HOT FIRE Pol EvaGreen HRM Mix (Solis BioDyne), 0.25 μM of forward and reverse primer and 20 ng DNA template. The solution was then overlaid with 15 μL PCR grade mineral oil (SIGMA). Amplification conditions included: 95°C for 15 min, then 45 cycles of 95°C for 30 s, 62°C for 30 s and 72°C for 15 s. Final hold temperatures were 95°C for 30 s and 25°C for 2 min. The products were then melted from 55°C to 95°C and melt curves assessed using the LightScanner (Idaho Technology Inc.).
SSR markers were screened and evaluated as described previously [20, 28, 61].
All mapping calculations were carried out in JoinMap V4  using the Kosambi function. Segregation and phase of all markers were checked and skewed markers (p < 0.05) were disregarded from further analysis. Linkage groups were formed using a maximum recombination fraction of 0.25 and a minimum LOD value of 7. The markers were then ordered using window size of 5 and a minimum LOD of 3. Rippling using a window size of 3 was used to visualize the marker order by both checking the minimum number of cross-overs and a maximum likelihood estimation for all possible orders. The linkage groups were then assigned a chromosome number based on the anchored SSR markers or markers that had been anchored using A. fistulosum - A. cepa monosomic addition lines , groups were visualized using Mapchart . QTL analysis was performed using RQTL . Using the framework map, a bin mapping set of 10 progeny was selected with minimization of expected bin size using the SAMPLEEXP command in MapPop .
Targeted marker development and Bin mapping
Sequences for the following transcription factor families were downloaded from ‘pfam’ : AP2, Dof, GRAS, HD, Myb, NAC, PHD, PLATZ, SET, Sigma70, WRKY, Whirly, BHLH, bZip, Arid and TCP. Translated assemblies of ‘CUDH2150’ transcriptome were searched for matches with these motifs using hmmsearch [67, 68] with E < 10-6 cutoff. SNP and indel variants identified in these contigs were filtered from GFF3 formatted read mapping output using Galaxy textual filtering tools. CAPS, indel and HRM markers were designed to these using Galaxy tools described in this paper. Markers were initially tested on parental and F1 samples and then on a bin mapping panel of 10 individuals. Markers were assigned to genetic map bins using MapPop 1.0 .