On raceway of the fish received a standard diet contaiing fishmeal and 10 percent fish oil, the experimental fish were fed a diet without fishmeal and protein source replaced with wheat gluten, corn gluten and soybean meal. The experimental diet also only used 5 percent fish oil and contained soybean oil in its place.
Growth protocol
One hundred and sixty females were mated to 80 males. After the eggs had eyed-up the family numbers were reduced to 95 full sib families and the same number of eggs from each family were pooled and mixed and then split into two separate lots of 4750 eggs. At first feeding the fry were separated into six 6000 l rearing tanks supplied with constant temperature (12 degree C) spring water. The fish were then transferred to 6 concrete rearing raceways when they averaged 25g and maintained in the raceways until they were harvested at an average weight of 600g.
Extracted molecule
total RNA
Extraction protocol
At an average harvest weight of approximately 600g, a random sample of 1032 fish from each diet group (344/raceway) was obtained. Each fish was weighed and measured for length and a fin clip was taken and stored in 100% ethanol for DNA extraction. Liver and muscle tissue collected for RNA isolation was placed in RNAlater (Ambion, Carlsbad, CA USA) and stored at 4°C for no longer than 3 weeks, according to the protocol recommended by Ambion. RNA was isolated using the TRIzol (Invitrogen, Carlsbad, CA USA) method. Briefly, tissue was first homogenized by placing a 5mm sterile, RNase-free, stainless steel bead into a round-bottom 2ml tube containing TRIzol and the tissue, then the tubes were placed in a Retsch MM300 Laboratory Vibration Mill for four min at 20 Hz. The protocol recommended by Invitrogen was followed for the rest of the RNA isolation.
Label
Cy5
Label protocol
Total RNA for each individual fish was isolated by the TRIzol method from liver and purified using Promega’s SV RNA Isolation Kit. cDNA was prepared and either Cy3- or Cy5-labeled using Superscript II Reverse Transcriptase (Invitrogen) and Genisphere’s Array 50 Kit. Each cDNA probe set was concentrated using Microcon YM-30 cut-off columns (Eppendorf), mixed with 100 μl SlideHyb 3 hybridization buffer (Ambion) and 1 μg human Cot-1 DNA (Invitrogen), incubated at 95°C for 10 minutes, then held at 45°C (hybridization temperature) until ready to use.
Channel 2
Source name
liver from large sized rainbow trout family fish on fish meal based diet
On raceway of the fish received a standard diet contaiing fishmeal and 10 percent fish oil, the experimental fish were fed a diet without fishmeal and protein source replaced with wheat gluten, corn gluten and soybean meal. The experimental diet also only used 5 percent fish oil and contained soybean oil in its place.
Growth protocol
One hundred and sixty females were mated to 80 males. After the eggs had eyed-up the family numbers were reduced to 95 full sib families and the same number of eggs from each family were pooled and mixed and then split into two separate lots of 4750 eggs. At first feeding the fry were separated into six 6000 l rearing tanks supplied with constant temperature (12 degree C) spring water. The fish were then transferred to 6 concrete rearing raceways when they averaged 25g and maintained in the raceways until they were harvested at an average weight of 600g.
Extracted molecule
total RNA
Extraction protocol
At an average harvest weight of approximately 600g, a random sample of 1032 fish from each diet group (344/raceway) was obtained. Each fish was weighed and measured for length and a fin clip was taken and stored in 100% ethanol for DNA extraction. Liver and muscle tissue collected for RNA isolation was placed in RNAlater (Ambion, Carlsbad, CA USA) and stored at 4°C for no longer than 3 weeks, according to the protocol recommended by Ambion. RNA was isolated using the TRIzol (Invitrogen, Carlsbad, CA USA) method. Briefly, tissue was first homogenized by placing a 5mm sterile, RNase-free, stainless steel bead into a round-bottom 2ml tube containing TRIzol and the tissue, then the tubes were placed in a Retsch MM300 Laboratory Vibration Mill for four min at 20 Hz. The protocol recommended by Invitrogen was followed for the rest of the RNA isolation.
Label
Cy3
Label protocol
Total RNA for each individual fish was isolated by the TRIzol method from liver and purified using Promega’s SV RNA Isolation Kit. cDNA was prepared and either Cy3- or Cy5-labeled using Superscript II Reverse Transcriptase (Invitrogen) and Genisphere’s Array 50 Kit. Each cDNA probe set was concentrated using Microcon YM-30 cut-off columns (Eppendorf), mixed with 100 μl SlideHyb 3 hybridization buffer (Ambion) and 1 μg human Cot-1 DNA (Invitrogen), incubated at 95°C for 10 minutes, then held at 45°C (hybridization temperature) until ready to use.
Hybridization protocol
Hybridization was carried out using a Tecan HS400 hybridization machine. GRASP 16k microarray slides [16] were prepared by washing for 30 seconds at 23°C with 0.2x SSC, 0.05% SDS, soaking in the same buffer for 2 minutes at 23°C, and washing for 1 minute at 45°C with 4x SSC, 0.05% SDS. Probes were injected onto slides and hybridization was carried out for 18 hours at 45°C with low agitation frequency. Slides were washed for 30 seconds and soaked for 1 minute at 43°C with 2x SSC, 0.1% SDS; this step was repeated two more times. Slides were then washed for 30 seconds and soaked for 1 minute at 25°C with 2x SSC; this step was repeated two more times. Slides were then washed for 30 seconds and soaked for 30 seconds at 25°C with 0.2x SSC, 0.05% SDS. Cy3 and Cy5 capture reagents (Genisphere) were prepared according to Genisphere’s protocol, and mixed with 50 μl Ambion SlideHyb 3 buffer and 50 μl nuclease-free H2O. The mixture was incubated at 80°C for 10 minutes, and then injected onto the slides. Hybridization was carried out for 3 hours, 30 minutes at 45°C with low agitation frequency. Slides were washed as described following first hybridization, and then dried for 2 minutes at 30°C.
Scan protocol
Scanning was performed using a Perkin Elmer ScanArray 5000.
Description
Comparison between the fishmeal and plantmeal diets within family M67/F143 Analyzed data from arrays A1Cy3-B1Cy5, A2Cy3-B2Cy5, A3Cy3-B3Cy5, and A4Cy3-B4Cy5 Comparison between diets for family M67/F143 Comparing expression in liver between fish from a large sized family on fishmeal to a fish from a medium sized family reared on a plant based diet Study's genotypes: family groups M67/F143 and M45/F85 fm.pm.fam1
Data processing
Microarray design and analysis- Genotyped samples linked to families that had 4 or more members sampled on each diet were identified and ranked according to average family weight on each diet. Size rankings corresponded to large (>750 g), medium (600-640 g), and small (<500 g). From these identified groups 2 families were chosen for microarray expression analysis that demonstrated individuals with growth differences between the two diets. Data for the two families used for analysis and slide arrangements are listed in table 3. The families compared for growth differences between diets used for this study only differed by one growth range, large fish on fish meal compared to medium fish on plant-based diet and medium sized fish on fishmeal diet compared to small fish on plant-based diet. This strategy to use size separation within family was an attempt to identify gene changes more specific to diet utilization than specific for growth differences. Array data pre-processing- The raw data intensities collected from two-channel arrays were pre-processed with scripts written in R language. These scripts included functions from the statistical packages MARRAY, ARRAYQUALITY and LIMMA from the Bioconductor project [81]. The array data pre-processing included arrays diagnostics, data quality assessment, data filtering, data background correction, and data normalization as outlined elsewhere. The scripts for data pre-processing are available on request (roger.vallejo@ars.usda.gov). Data filtering- The raw data intensities generated by the ScanArray Express software (Perkin Elmer, MA) had a column flag with values of 1= spot not found, 3= good spot, and 4= bad spot. Spots with flags of 1 and 4 were filtered out, in addition to those with columns Name and ID that had these remarks: Name = “Empty,” “Blank Well,” “EmptySpot” and ID = “NAC.”Background correction- For background correction to assess differential expression, we used the method “normexp” with offset = 50 implemented in the computer package LIMMA, which was often found to be more robust than a simple background subtraction when using outputs from most image analysis programs. This method adjusts the foreground adaptively for the background intensities and results in strictly positive adjusted intensities. The use of this offset allows for the filtering out of gene expression data with signal intensities <50. The effect of using this background adjustment method is to stabilize the variability of the M values [i.e., M = log 2(Cy5) - log 2(Cy3)] as a function of signal intensity. Data normalization. A within-array or two-channel normalization method was used to perform direct comparisons of RNA samples hybridized in the same array. The “robust spline normalization” method of data normalization was used. Statistical analysis of microarray data-The statistical analysis of pre-processed microarray data was performed with scripts written in R language. These scripts included functions from the statistical package LIMMA from the Bioconductor project [81]. The scripts for statistical data analysis are available on request (roger.vallejo@ars.usda.gov). The study design included two treatment factors with two levels each: diet and family. The focus of this study was to identify differentially expressed genes (DEGs) due to diet effects within each family and across families. We were also interested in identifying DEGs due to family effects within each diet and across diets. Therefore, linear models were fitted for each of these four contrasts (family 1 vs. family 2 / diet 1; family 1 vs. family 2 / diet 2; diet 1 vs. diet 2 / family 1; and diet 1 vs. diet 2 / family 2) to identify genes that had significant differential gene expression for “treatment” effects. By study design restrictions, we could model the “dye” effect in the fitted linear models only when comparing families within each diet. The basic statistic used for significance analysis is the moderated t-statistic, which is computed for each probe and for each contrast. This has the same interpretation as an ordinary t-statistic except that the standard errors have been moderated across genes (i.e., shrunk toward a common value) with a simple Bayesian model. The adjusted P values (aka q values) are an estimate of the false discovery rate (FDR). We calculated the FDR by the Benjamini and Hochberg method. For all hybridizations, all features with a a FDR of <0.2 were considered to be significantly different. A within-array or two-channel normalization method was used to perform direct comparisons of RNA samples hybridized in the same array. The “robust spline normalization” method of data normalization was used. Statistical analysis of microarray data-The statistical analysis of pre-processed microarray data was performed with scripts written in R language. These scripts included functions from the statistical package LIMMA from the Bioconductor project.