Deep Learning-Based Fish Detection Using Above-Water Infrared Camera for Deep-Sea Aquaculture: A Comparison Study.

Long-term, automated fish detection provides invaluable data for deep-sea aquaculture, which is crucial for safe and efficient seawater aquafarming. In this paper, we used an infrared camera installed on a deep-sea truss-structure net cage to collect fish images, which were subsequently labeled to establish a fish dataset. Comparison experiments with our dataset based on Faster R-CNN as the basic objection detection framework were conducted to explore how different backbone networks and network improvement modules influenced fish detection performances. Furthermore, we also experimented with the effects of different learning rates, feature extraction layers, and data augmentation strategies. Our results showed that Faster R-CNN with the EfficientNetB0 backbone and FPN module was the most competitive fish detection network for our dataset, since it took a significantly shorter detection time while maintaining a high AP50 value of 0.85, compared to the best AP50 value of 0.86 being achieved by the combination of VGG16 with all improvement modules plus data augmentation. Overall, this work has verified the effectiveness of deep learning-based object detection methods and provided insights into subsequent network improvements.

Responses of microeukaryotic community structure to a Phaeocystis globosa bloom in a semi-enclosed subtropical bay.

The occurrence of Phaeocystis globosa, a harmful algal bloom species in Chinese coastal waters, has significant impacts on marine organisms and poses a threat to the safety of coastal nuclear power plants. Although previous studies have established a close association between P. globosa blooms and the bacterial community, the relationship between the microeukaryotic community and P. globosa blooms remains poorly understood. In this study, the variations in the microeukaryotic community resulting from a P. globosa bloom were analyzed using 18S rRNA gene amplicon sequencing. The results indicated that the diversity of the microeukaryotic community during the bloom phase was significantly higher than that during the dissipation phase. The microeukaryotic community compositions varied significantly between the two phases of the P. globosa bloom. During the bloom phase, the dominant microeukaryotic was Viridiplantae, which was then replaced by Dinoflagellata during the dissipation phase. Co-occurrence network analysis showed that the relationship among the microeukaryotic community during the bloom phase was more complex than that during the dissipation phase, and the keystone taxa varied as the bloom progressed. Additionally, microeukaryotic community assembly was primarily driven by stochastic processes during the bloom phase based on the ß-nearest taxon distance, whereas it was driven by both deterministic processes and stochastic processes during the dissipation phase. Overall, our findings provide novel insight into the mechanisms and interactions involved in microeukaryotic community dynamics in environments disturbed by P. globosa blooms.

Signal Transduction Mechanisms for Glucagon-Induced Somatolactin Secretion and Gene Expression in Nile Tilapia (Oreochromis niloticus) Pituitary Cells.

Glucagon (GCG) plays a stimulatory role in pituitary hormone regulation, although previous studies have not defined the molecular mechanism whereby GCG affects pituitary hormone secretion. To this end, we identified two distinct proglucagons, Gcga and Gcgb, as well as GCG receptors, Gcgra and Gcgrb, in Nile tilapia (Oreochromis niloticus). Using the cAMP response element (CRE)-luciferase reporter system, tilapia GCGa and GCGb could reciprocally activate the two GCG receptors expressed in human embryonic kidney 293 (HEK293) cells. Quantitative real-time PCR analysis revealed that differential expression of the Gcga and Gcgb and their cognate receptors Gcgra and Gcgrb was found in the various tissues of tilapia. In particular, the Gcgrb is abundantly expressed in the neurointermediate lobe (NIL) of the pituitary gland. In primary cultures of tilapia NIL cells, GCGb effectively stimulated SL release, with parallel rises in the mRNA levels, and co-incubation with the GCG antagonist prevented GCGb-stimulated SL release. In parallel experiments, GCGb treatment dose-dependently enhanced intracellular cyclic adenosine monophosphate (cAMP) accumulation with increasing inositol 1,4,5-trisphosphate (IP3) concentration and the resulting in transient increases of Ca2+ signals in the primary NIL cell culture. Using selective pharmacological approaches, the adenylyl cyclase (AC)/cAMP/protein kinase A (PKA) and phospholipase C (PLC)/IP3/Ca2+/calmodulin (CaM)/CaMK-II pathways were shown to be involved in GCGb-induced SL release and mRNA expression. Together, these results provide evidence for the first time that GCGb can act at the pituitary level to stimulate SL release and gene expression via GCGRb through the activation of the AC/cAMP/PKA and PLC/IP3/Ca2+/CaM/CaMK-II cascades.

Irisin stimulates gonadotropins gene expression in tilapia (Oreochromis niloticus) pituitary cells.

The link between energy metabolism and reproduction is well known in vertebrates. Irisin, the product of fibronectin type III domain-containing protein 5 (FNDC5) gene, plays an important role in energy homeostasis. However, biological actions of irisin on reproduction remain elusive. To address this gap, we examined the direct effects of irisin on luteinizing hormone ß (LHß) and follicle-stimulating hormone ß (FSHß) gene expression in tilapia pituitary cells. As a first step, the transcripts of FNDC5 were detected in the proximal pars distalis (PPD), but not in the rostral pars distalis (RPD) and neurointermediate lobe (NIL) of the tilapia pituitary by RT-PCR. In the tilapia pituitary, irisin immunoreactive signals were also detected in PPD region. In primary cultures of tilapia pituitary cells, irisin was effective in stimulating both LHß and FSHß mRNA levels in vivo and in vitro. In cultured pituitary cells of tilapia, removal of endogenous irisin by immunoneutralization using irisin antiserum inhibited LHß and FSHß gene expression. Salmon gonadotrophin releasing hormone (sGnRH) increased LHß and FSHß mRNA levels in tilapia pituitary but these stimulatory actions were not either enhanced by treatment with irisin or blocked by irisin antiserum. Furthermore, the stimulation on LHß and FSHß mRNA expression was coincident with the enhancement of LHß and FSHß mRNA stability after irisin treatment. These results provide evidence that irisin may serve as a novel intrapituitary factor maintaining gonadotropins gene expression in tilapia pituitary.

Irisin inhibition of growth hormone secretion in cultured tilapia pituitary cells.

Irisin, the product of fibronectin type III domain-containing protein 5 (FNDC5) gene, is well-documented to be a regulator of energy metabolism. At present, not much is known about its biological function in non-mammalian species. In this study, a full-length tilapia FDNC5 was cloned and its tissue expression pattern has been confirmed. Based on the sequence obtained, we produced and purified recombinant irisin which could induce uncoupling protein 1 (UCP1) gene expression in tilapia hepatocytes. Further, the rabbit polyclonal irisin antiserum was produced and its specificity was confirmed by antiserum preabsorption. In tilapia pituitary cells, irisin inhibited growth hormone (GH) gene expression and secretion and triggered rapid phosphorylation of Akt, Erk1/2, and p38 MAPK. Furthermore, irisin-inhibited GH mRNA expression could be prevented by inhibiting PI3K/Akt, MEK1/2, and p38 MAPK, respectively. Apparently, fish irisin can act directly at the pituitary level to inhibit GH transcript expression via multiple signaling pathways.

Dopamine inhibits somatolactin gene expression in tilapia pituitary cells through the dopamine D2 receptors.

Dopamine (DA) is an important neurotransmitter in the central nervous system of vertebrates and possesses key hypophysiotropic functions. Early studies have shown that DA has a potent inhibitory effect on somatolactin (SL) release in fish. However, the mechanisms responsible for DA inhibition of SL gene expression are largely unknown. To this end, tilapia DA type-1 (D1) and type-2 (D2) receptor transcripts were examined in the neurointermediate lobe (NIL) of the tilapia pituitary by real-time PCR. In tilapia, DA not only was effective in inhibiting SL mRNA levels in vivo and in vitro, but also could abolish pituitary adenylate cyclase-activating polypeptide (PACAP)- and salmon gonadotropin-releasing hormone (sGnRH)-stimulated SL gene expression at the pituitary level. In parallel studies, the specific D2 receptor agonists quinpirole and bromocriptine could mimic the DA-inhibited SL gene expression. Furthermore, the D2 receptor antagonists domperidone and (-)-sulpiride could abolish the SL response to DA or the D2 agonist quinpirole, whereas D1 receptor antagonists SCH23390 and SKF83566 were not effective in this respect. In primary cultures of tilapia NIL cells, D2 agonist quinpirole-inhibited cAMP production could be blocked by co-treatment with the D2 antagonist domperidone and the ability of forskolin to increase cAMP production was also inhibited by quinpirole. Using a pharmacological approach, the AC/cAMP pathway was shown to be involved in quinpirole-inhibited SL mRNA expression. These results provide evidence that DA can directly inhibit SL gene expression at the tilapia pituitary level via D2 receptor through the AC/cAMP-dependent mechanism.

Adropin induction of lipoprotein lipase expression in tilapia hepatocytes.

The peptide hormone adropin plays a role in energy homeostasis. However, biological actions of adropin in non-mammalian species are still lacking. Using tilapia as a model, we examined the role of adropin in lipoprotein lipase (LPL) regulation in hepatocytes. To this end, the structural identity of tilapia adropin was established by 5'/3'-rapid amplification of cDNA ends (RACE). The transcripts of tilapia adropin were ubiquitously expressed in various tissues with the highest levels in the liver and hypothalamus. The prolonged fasting could elevate tilapia hepatic adropin gene expression, whereas no effect of fasting was observed on hypothalamic adropin gene levels. In primary cultures of tilapia hepatocytes, synthetic adropin was effective in stimulating LPL release, cellular LPL content, and total LPL production. The increase in LPL production also occurred with parallel rises in LPL gene levels. In parallel experiments, adropin could elevate cAMP production and up-regulate protein kinase A (PKA) and PKC activities. Using a pharmacological approach, cAMP/PKA and PLC/inositol trisphosphate (IP3)/PKC cascades were shown to be involved in adropin-stimulated LPL gene expression. Parallel inhibition of p38MAPK and Erk1/2, however, were not effective in these regards. Our findings provide, for the first time, evidence that adropin could stimulate LPL gene expression via direct actions in tilapia hepatocytes through the activation of multiple signaling mechanisms.