Dietary tryptophan overdose in Nile tilapia (Oreochromis niloticus) juveniles: growth performance, hematology, fillet traits, fatty acid profile, lipidome and gene expression
Description
This dataset contains the raw data of a 20-day feeding trial in which Nile tilapia (Oreochromis niloticus) juveniles received either a control diet or the same diet with tryptophan at approximately eight times the recommended level ([fill: exact inclusion, g/kg]). Twelve hapas (six per treatment) were stocked with seven fish each, and the hapa was the experimental unit. The data include growth performance, feed intake, feed conversion, fillet weight and yield, mortality and survival (hapa means); fillet CIELAB color, pH and proximate composition; erythrocyte parameters (RBC, hematocrit and MCV); fillet fatty acid profile and the fatty acid profile of the diet; Ct values of qPCR for beta-actin, pla2, hsp70, fads2, scd, acca, cpt1 and acox1 in fillet; and untargeted LC-MS lipidomics of the fillet (148 annotated features, of which 136 lipids were retained after removal of contaminants and non-lipid metabolites). Hematology, fatty acids, qPCR and lipidomics were measured in the same fish (one per hapa). The data are provided as one Excel workbook (one sheet per analysis, a README sheet and a Variables sheet describing every variable and its unit) and as CSV files with the same content. The Python script supplied with the data reproduces all statistical analyses, tables and figures of the related article.
Files
Steps to reproduce
Install Python 3.10 or later and the packages pandas, numpy, scipy, matplotlib and openpyxl (pip install pandas numpy scipy matplotlib openpyxl). Place the script and the workbook in the same folder and run: python tryptophan_tilapia_analysis.py tryptophan_tilapia_dataset.xlsx. The script uses the hapa as the experimental unit (means of the fish sampled per hapa). For each variable, normality was checked with the Shapiro–Wilk test in each group and homogeneity of variances with the Levene test (median); groups were compared with Student t test, Welch t test or two-sided Mann–Whitney test accordingly. Relative gene expression was calculated by the 2^-ΔΔCt method with beta-actin as reference gene in each run, and tests were applied to ΔCt values. Lipid peak areas were normalized to the total lipid signal (zeros replaced by half of the minimum value of each lipid), log2-transformed and analyzed with a linear model including treatment and analytical batch; p values were adjusted by the Benjamini–Hochberg procedure, and the overall lipidome was compared by PERMANOVA (Euclidean distance on autoscaled data, exact permutation test).