Automated Measurement of Ovary Development in Atlantic Salmon Using Deep Learning.

OBJECTIVE: Salmon breeding companies control the egg stripping period through environmental change, which triggers the need to identify the state of maturation. Ultrasound imaging of the salmon ovary is a proven non-invasive tool for this purpose; however, the process is laborious, and the interpretation of the ultrasound scans is subjective. Real-time ultrasound image segmentation of Atlantic salmon ovary provides an opportunity to overcome these limitations. However, several application challenges need to be addressed to achieve this goal. These challenges include the potential for false-positive and false-negative predictions, accurate prediction of attenuated lower ovary parts and resolution of inconsistencies in predicted ovary shape. METHODS: We describe an approach designed to tackle these obstacles by employing targeted pre-training of a modified U-Net, capable of performing both segmentation and classification. In addition, a variational autoencoder (VAE) and generative adversarial network (GAN) were incorporated to rectify shape inconsistencies in the segmentation output. To train the proposed model, a data set of Atlantic salmon ovaries throughout two maturation periods was recorded. RESULTS: We then tested our model and compared its performance with that of conventional and novel U-Nets. The method was also tested in a salmon on-site ultrasound examination setting. The results of our application indicate that our method is able to efficiently segment salmon ovary with an average Dice score of 0.885 per individual in real-time. CONCLUSION: These results represent a competitive performance for this specific application, which enables us to design an automated system for smart monitoring of maturation state in Atlantic salmon.

Ultrasound as a noninvasive tool for monitoring reproductive physiology in male Atlantic salmon (Salmo salar).

We examined the potential for ultrasound as a noninvasive tool for maturation monitoring in Atlantic salmon (Salmo salar) males. Ultrasound examination and measurements were compared to common practices for maturation monitoring such as gonadosomatic index (GSI), sex hormone analysis, and histological analysis of spermatogenesis. There were significant correlations (R2  = 0.68, P < 0.01) between ultrasound-based measurements of the left testis and total testes weight and GSI, and ultrasound could be used for noninvasive GSI measurements. Echogenicity of ultrasound images corresponded to the histological stages observed, which added nuance to ultrasound-based GSI measurements during final weeks preceding stripping. We propose that ultrasound can be used as an alternative to more invasive methods for sexual maturation monitoring in wild and farmed Atlantic salmon males. Using ultrasound technology, we have established a quick and noninvasive method that could reduce the number of stressful handlings and unwanted sacrifice of broodfish required for maturation monitoring in Atlantic salmon males.

Ultrasound as a noninvasive tool for monitoring reproductive physiology in female Atlantic salmon (Salmo salar).

Aiming to explore ultrasound technology as a noninvasive method for maturation monitoring, we compared ultrasound observations and measurements in female Atlantic salmon (Salmo salar) during the last year before ovulation with standard, invasive methods such as gonadosomatic index (GSI), gonad histology and sex hormone analysis. Ultrasound measurements of ovaries correlated strongly (R > 0.9, P < 0.01) with ovary weight and GSI, and could be used as a noninvasive tool for GSI estimation. Using ultrasound, we were able to identify females with advanced oocyte development and elevated sex hormone and GSI levels earlier than previously observed. Histological studies confirmed these observations showing oocyte yolk accumulation 10 months before ovulation and 8 months before significant increase in sex hormones. Levels of the sex hormone 11-keto testosterone (11-KT) indicated a new role of this hormone at final maturation in salmon females. We propose the use of ultrasound as an alternative method to traditionally used invasive methods during sexual maturation monitoring in wild and farmed Atlantic salmon broodstock populations. Eliminating sacrifice of valuable broodfish, and reducing handling stress, would improve animal welfare in present-day broodstock management.

Resolving the complexity of vitellogenins and their receptors in the tetraploid Atlantic salmon (Salmo salar): Ancient origin of the phosvitin-less VtgC in chondrichthyean fishes.

Egg yolk proteins are mainly derived from vitellogenin (Vtg), and serve as essential nutrients during early development in oviparous organisms. Vertebrate Vtgs are predominantly synthesized in the liver of maturing females, and are internalized by the oocyte after binding to specific surface receptors (VtgR). Here, we clarify the evolutionary history of vertebrate Vtgs, including the teleost VtgC, which lacks phosvitin, and investigate the repertoire of Vtgs and VtgRs in the tetraploid Atlantic salmon (Salmo salar). Conserved synteny of the vtg genes in elephant fish (Callorhinchus milii) strongly indicates that the vtg gene cluster was present in the ancestor of tetrapods and ray-finned fish. The shortened phosvitin in the VtgC ortholog of this chondrichthyean fish may have resulted from early truncation events that eventually allowed the total disappearance of phosvitin in teleost VtgC. In contrast, the tandem-duplicated VtgCs identified in the spotted gar (Lepisosteus oculatus) both contain the phosvitin domain. The Atlantic salmon genome harbors four vtg genes encoding the complete VtgAsa1, phosvitin-less VtgC, and truncated VtgAsb proteins; vtgAsa2 is a pseudogene. The three vtg genes were mainly expressed in the liver of maturing females, and the vtgAsa1 transcript predominated prior to spawning. The splice variant lacking the O-linked sugar domain dominated ovarian expression of vtgr1 and vtgr2. Strongly increased vtgAsa1 expression during vitellogenesis contrasted with the peaks of vtgr1 and vtgr2 in the previtellogenic oocytes, which gradually decreased over the same period. Recycling of the oocyte VtgRs is probably not sufficient to maintain receptor number during vitellogenesis.