Skip to main content
European Commission logo print header

Microarray analysis of salmon transcriptomes: evaluation of a large cDNA array and production of a targeted diagnostic oligonucleotide array

Final Activity Report Summary - MAST (Microarray analysis of salmon transcriptomes: Evaluation of a large cDNA array and production of a targeted diagnostic oligonucleotide array)

The current use of fish oils (FO), from feed-grade fisheries, in aquaculture feeds is not sustainable and will constrain continuing growth of aquaculture. This had led to research investigating the substitution of FO with plant seed (vegetable) oils (VO). VOs can be rich in C18 polyunsaturated fatty acids but are devoid of the omega-3 highly unsaturated fatty acids (HUFA) that are abundant in FO and essential for human beings. Changes in salmon lipid and fatty acid metabolism induced by the replacement of FO by VO can alter fatty acid composition of the flesh and then negatively affect the nutritional quality of the fish. After advanced training in transcriptomics, including the extraction and purification of RNA, labelling of RNA, performing hybridisations and analyses of arrays, the fellow performed two validations of the TRAITS/SGP salmon cDNA microarray. The first validation utilised samples taken at four different time points from Atlantic salmon fed diets containing FO or a VO blend. The second validation involved samples from salmon fed with different single VOs or FO. In both cases, more than 100 genes mainly linked to lipid and fatty acid pathways were significantly up-regulated and some of them were subjected to quantitative PCR (Q-PCR) for further validation. These experiments confirmed that the TRAITS/SGP cDNA microarrays were a highly effective and efficient research tool.

The TRAITS/SGP cDNA microarray was originally envisaged as an intermediate tool to identify genes and gene pathways involved in controlling/regulating a number of commercially important traits in Atlantic salmon. Following identification of such genes a second, more targeted, oligonucleotide (oligo) array was fabricated for more intensive studies. To validate the oligo array, a subset of the same samples interrogated with the cDNA array were hybridised to the oligo array, with the expectation that similar expression patterns would be observed. The samples came from fish fed FO or the VO blend. The dye-swap technique was used for the hybridisations. Analysis indicated that c. 70% of the features identified by the cDNA array as being differentially regulated by diet were also identified as significant responders by oligo array analysis. This figure mirrors that found by the other TRAITS / SGP partners for their traits/genes of interest.

This project also aimed to identify salmon candidate genes associated with the ability to synthesis and/or retain n-3 HUFA in the flesh, and thus provide the genetic basis for new marker-assisted selection strategies. Therefore, the fellow was involved in analysing samples from another EU project, the FP6 IP AQUAMAX (016249-2). Fifty families had been grown for 3 months on a VO diet and the families phenotyped for flesh n-3HUFA level. Eight families were identified as having significantly different flesh n-3HUFA levels and they were selected for liver transcriptome analysis. Specifically, these 8 families were the most extreme ones, 4 of them exhibiting a high flesh level of n-3HUFA and the 4 other ones showing a low flesh level of n-3HUFA. The candidate genes involved in fatty acid pathways did not show any significant difference between the families exhibiting a high flesh level of n-3HUFA and those with a low flesh level of n-3HUFA when studied by Q-PCR.

Transcriptomic analysis using the cDNA array confirmed these genes were not significantly responding but some other genes involved in the folding and transport of lipoproteins significantly responded. Unfortunately, there was insufficient time for the fellow to study these responding genes by Q-PCR so they will be further investigated by the host institute in the following months. Thus, these results seem to indicate that there may be differences in gene expression that could contribute to the difference in n-3HUFA levels in flesh observed between the families. Several manuscripts are in preparation or submitted.