A wheat SNP chip for the study of homoeologous gene expression in hexaploid wheat

We propose to use recently developed technologies and sequence information to create and trial a novel transcriptomics-based resource for the wheat community. This resource, which will be fully upgradable without further cost to the community, will consist of a wheat array capable of monitoring the expression of up to 90,000 different homoeolog and paralog transcripts. This approach will remove many of the seen and unseen problems associated with the existing GeneChip and EST-based platforms and will be a valuable source of new and novel information on the formation of Triteace polyploids and the differential expression of the three different genomes that make up allohexaploid wheat. We propose to open up this resource to the entire wheat community as an unencumbered facility free of MTAs and follow on IP agreements.

Analysis of the wheat transcriptome can now be undertaken using an Affymetrix GeneChip or numerous academic/public oligo and EST spotted microarrays. As part of work designed to merge our EST array and GeneChip datasets we have carried out controlled experiments to compare and contrast the two systems. This comparison showed that rather than helping to accurately describe the wheat transcriptome, the current transcriptomics-based resources are at best confusing and at worst totally misleading: The wheat Affymetrix GeneChip contains 13,605 Affymetrix features matching one or more EST array features on the Bristol EST array and the Bristol array contains 6,315 ESTs with one or more matches on the Affymetrix GeneChip. However, when we used both platforms to analysis the transcriptome of the same wheat samples we found that unless we only selected highly expressed genes we could not make a global comparison of the common expressed sequences. To investigate the possibility that one platform was working while the other was not, we used Q-RT-PCR; unfortunately use of Q-RT-PCR only confused the issue further with these results often disagreeing with the data from either the Affymetrix GeneChip or EST-arrays (for ~ 40% of cases). Such discrepancies have been seen before, for instance in a recent study on maturing and germinating wheat grain we were only able to confirm via Q-RT-PCR approximately 70% of our EST array observations. These results were in agreement with the findings of other groups undertaking similar studies. In our previous work we suggested that these results might be due to both variation inherent to microarray experiments and the inability of arrays to differentiate between the hybridisation of closely related sequences (homoeologs and paralogs). To overcome this limitation we now proposal to design a second-generation oligo array capable of discriminating between homoeologs and paralogs.