Selection of Nonlethal Early Biomarkers to Predict Gilthead Seabream (Sparus aurata) Growth

Abstract

One of the main challenges in aquaculture is the constant search for sustainable alternative feed ingredients that can successfully replace fishmeal (FM) without any negative effects on fish growth and health. The goal of the present study was to develop a toolbox for rapidly anticipating the dynamics of fish growth following the introduction of a new feed; nonlethal, biochemical, and molecular markers that provide insights into physiological changes in the fish. A nutritional challenge by feeding a conventional feed rich in FM protein (FM diet) versus an experimental feed rich in plant protein (PP) and low FM inclusion (PP diet), in 20 different families of gilthead sea bream (Sparus aurata) was performed. Fifteen and 30 days after the initiation of the nutritional challenge, the transcriptional response of gilthead seabream erythrocytes along with classical hematological biochemical markers were compared. Zootechnical, biochemical, and transcriptome data from each family under different dietary treatments were combined into a classification model to identify variables that can predict the growth rate at the end of the 14-month farming period (July 2018-September 2019). A highly accurate model was produced (A > 80%) based on the combination of seven markers (five molecular and two biochemical markers) and with high potential in separating faster and slower growing fish as early as 30 days after the initiation of feeding.

Keywords: biochemical markers; growth potential; nonlethal markers; transcriptome markers.

Figure 1

Comparison of periodical family SGR (median) between the two diets. (A) SGR September–November, (B) SGR November–January, (C) SGR January–March, (D) SGR March–July, and (E) SGR July–August. Light blue and green denote the families that are fed on the FM diet and the PP diet, respectively. Significance levels are presented on the plot (⁣^∗p < 0.05, ⁣^∗∗p < 0.01, ⁣^∗∗∗p < 0.001, ⁣^∗∗∗∗p < 0.0001).

Figure 2

Muscle fat content measured at the end of the experiment. Light blue and green denote the families that are fed on the FM diet and the PP diet, respectively.

Figure 3

Body weight of all families at the end of the trial. Light blue and green denote the families that are fed on the FM diet and the PP diet, respectively. Significance levels are presented on the plot (⁣^∗p < 0.05, ⁣^∗∗p < 0.01, ⁣^∗∗∗p < 0.001, and ⁣^∗∗∗∗p < 0.0001).

Figure 4

Fat content to body weight ratio of all families at the end of the trial. Light blue and green denote the families that are fed on the FM diet and the PP diet, respectively. Significance levels are presented on the plot (⁣^∗p < 0.05, ⁣^∗∗p < 0.01, ⁣^∗∗∗p < 0.001, and ⁣^∗∗∗∗p < 0.0001).

Figure 5

Triglycerides levels in blood serum (A) 15 days (D15) and (B) 30 days (D30) following the start of the feeding trial. Higher values were observed in the PP diet group 30 days after trial initiation. Significance levels are presented on the plot (⁣^∗p < 0.05, ⁣^∗∗p < 0.01, ⁣^∗∗∗p < 0.001, and ⁣^∗∗∗∗p < 0.0001).

Figure 6

Cholesterol levels in blood serum (A) 15 days (D15) and (B) 30 days (D30) following the start of the feeding trial. Significance levels are presented on the plot (⁣^∗p < 0.05, ⁣^∗∗p < 0.01, ⁣^∗∗∗p < 0.001, and ⁣^∗∗∗∗p < 0.0001).

Figure 7

Volcano plots depicting genes upregulated in both dietary groups on the two sampling days: (A) D15 and (B) D30.

Figure 8

Protein content in blood serum (A) 15 days (D15) and (B) 30 days (D30) following the start of the feeding trial. Significance levels are presented on the plot (⁣^∗p < 0.05, ⁣^∗∗p < 0.01, ⁣^∗∗∗p < 0.001, and ⁣^∗∗∗∗p < 0.0001).

Figure 9

GO term enrichment analysis of the common DEGs in the three GO annotation domains: biological processes (BPs), cell components (CCs), and molecular functions (MFs).

Figure 10

Random forest classifier. Graph shows the molecular and biochemical markers correlated with final growth in a hierarchical manner. Superscripts derive from the multiple regression analysis.