Haemal lordosis is a frequent abnormality of the vertebral column. It has been recorded to develop in different finfish species, during the hatchery rearing phase. Under certain conditions, this abnormality reaches a high prevalence and severity degree, with significant effects on the external morphology of the fish. We show that haemal lordosis recovers during the on-growing of Gilthead seabream in sea cages. At the end of the hatchery phase, 1700 seabream juveniles were tagged electronically and examined for the presence of haemal lordosis. Subsequently, their morphology was examined periodically up to the end of the on-growing period. We found that the prevalence of fish with a lordotic external morphology decreased during the studied period by approximately 50%. Interestingly, 27% of the recovered fish presented a completely normal vertebral column. Geometric morphometric analysis showed no significant differences in the body shape between the fish with a recovered normal phenotype and the fish that were normal since the beginning of the on-growing period. Our results provide the first evidence for the recovery of lordosis during the growth of fish. A mechanism with multiple levels of remodeling of abnormal bones is suggested.
The presence of skeletal abnormalities in reared fish is a major problem of product quality, with significant effects on fish morphology and biological performance1. Although early sort-out of the abnormal juveniles minimizes the negative effects of abnormalities on the quality of the final product, this leads to a significant financial loss for hatcheries (e.g. over 50 million € per year for the Mediterranean aquaculture1). In most of the species, skeletal abnormalities develop during the embryonic, larval and early juvenile stages2,3,4,5,6, due to unfavorable rearing conditions7,8,9,10,11, genetic factors12,13,14 and/or due to occasional accidental deviations from the standard operating procedures. Existing literature suggests that prevalence of skeletal abnormalities might be decreased with the growth of the fish, as a result of the high lethality of some deformity types (e.g. prehaemal kyphosis15; haemal vertebral compression and fusion3). Recovery of skeletal abnormalities has only been recorded for the case of gill-cover abnormalities of light severity in Gilthead seabream16 and Atlantic salmon17. To our knowledge, there is no evidence on the recovery of vertebral abnormalities in finfish.
Vertebral lordosis (V-shape curvature) is a frequent abnormality of the reared finfish. Depending on the abnormality severity, the effects of lordosis on the external morphology display a continuous distribution, ranging from insignificant to severe body-shape alterations18. In physoclistous fish, lordosis was initially shown to affect the prehaemal vertebrae and arise from the failure of swimbladder inflation19,20. As it was later reported in Dicentrarchus labrax21 and Pagrus major22, lordosis may also develop in fish with a normally inflated swimbladder, on the haemal part of the vertebral column, as a result of high swimming activity. Nowadays, prehaemal lordosis is considered as a solved problem for finfish hatcheries, through mainly the application of appropriate methodology for the successful inflation of the swimbladder1. Concerning haemal lordosis, existing literature suggests that controlling the swimming intensity of the early juveniles might not completely solve the problem, since more factors have been shown to be significantly involved in its development (e.g. developmental temperature10,23; retinol levels in the larval diet24; genetic background12).
We examined whether haemal lordosis is further evolving during the growth of fish, in terms of its prevalence and effects on body shape. To reach this goal, we electronically tagged a juvenile population of Gilthead seabream (Sparus aurata L.) at the end of the hatchery phase and we followed fish morphology up to the end of the on-growing period. Selected fish is a major species of the European aquaculture, with a production of 161 thousand tons and 700 million juveniles in 2016 (source: FEAP).
Materials and Methods
Experimental design and fish origin
One thousand seven hundred seabream juveniles (86 ± 7 mm standard length, SL) were tagged electronically (FDX-B, Trovan Ltd, USA). Fish morphology was examined periodically until the end of the on-growing period (commercial size, 262 ± 14 mm SL) (Table 1). During each sampling, all specimens were anaesthetized by bath immersion (ethyleneglycol-monophenylether, Merck, 0.2–0.5 mL L−1), photographed on the left side and scanned for ID recognition before returning to the sea cage. All photographs were taken by means of a Canon PowerShot G9 camera, mounted on a tripod and positioned perpendicularly to the specimens. In each sampling period, missing fish were recorded. The significance of the differences in mortality rate between the two groups of fish (i.e. lordotic and normal) at each sampling age was tested by G-test25.
The examined group of fish originated from a common larval population and egg batch. Fish were reared according to the standard methodology followed by commercial hatcheries and cage farms for juvenile production and on-growing respectively26. At the end of the hatchery phase, 2000 juveniles (6.1 ± 1.9 g mean weight, 137 days post-hatching, dph) were transported to a commercial sea cage (6 m length, 6 m width, 8 m depth) for on-growing. Larvae were reared at an initial stocking density of 100 individuals per litre, in the presence of background phytoplankton (Chlorella sp.), with feeding on rotifers Brachionus plicatilis (5–27 dph), Artemia sp. nauplii (17–40 dph), and finally on inert commercial diets (>27 dph). Water temperature was maintained between 19–21 °C during the hatchery period, whereas during on-growing period it followed the natural seasonal fluctuations (Fig. S1). Rearing of Gilthead seabream embryos, larvae and juveniles was performed under routine production conditions at Andromeda S.A. This company is registered (registration number GGN 5200700699992) for aquaculture production in Greece and has secured a GLOBALG.A.P quality certification, which requires a certified Veterinary Doctor to periodically verify fish health and welfare. Animal sampling followed routine procedures and samples were collected by a qualified member of staff from a standard production cycle. The legislation and measures implemented by the commercial producer complied with existing Greek (PD 56/2013) and EU (Directive 63/2010) legislation (protection of animals kept for farming). Production and sampling, by an experienced worker, were optimised to avoid unnecessary pain, suffering or injury and to maximise larval survival.
Categorization of the external morphology
Morphological examination of lordosis was carried out by three independent observers and it was blind in respect to fish pit-tag number. Discrimination of the fish with lordotic external morphology from those with a normal one was based on the dorsal shift of the caudal peduncle in the abnormal fish23 (Fig. 1A,B). Fish without a typical normal or lordotic external morphology were categorized as of uncertain phenotype (Fig. 1C).
To validate the methodology followed for the examination of external morphology, a random sample of 100 juveniles was taken from the source population, simultaneously to the tagging of experimental fish. Sampled juveniles were anaesthetized, photographed and stored at −20 °C. Based on their external morphology, juveniles were scored as normal, lordotic or of uncertain phenotype (Fig. 1). Subsequently, the results of phenotypic scoring were radiographically validated.
Based on the evolution of external morphology during the studied period, 146 adults were taken at 434 days post-tagging (dpt) and radiographically examined to further analyze and verify the observed external morphology. Radiographed adults were representative samples of fish with different external phenotypes (Table 2). All samples collected for x-raying, were euthanatized with an overdose of anesthetic. Radiography was performed with 50 KV voltage, 400 mA intensity and 0.002 s exposure time (Econet medical PpP-60HF, AGFA CR10).
Differences in body shape between juveniles (1 day post-tagging, dpt) with a normal, lordotic or uncertain external morphology were examined by geometric morphometry. Analysis included all the 174 juveniles of lordotic and uncertain external morphology (Table 2), together with a random sample of 150 juveniles with normal morphology. A similar analysis was performed on the 71 lordotic fish and the 17 fish of uncertain morphology at the final sampling (434 days post-tagging, Table 2), together with a random sample of 50 fish with normal morphology.
Differences in body shape between adults that recovered from lordosis and those with a normal external morphology since the beginning of the trial were examined by geometric morphometry. Analysis included all the x-rayed adults with recovered phenotype and 25 randomly selected and x-rayed normal adults (Table 2). All the x-rayed adults with a severe lordosis were also included in the study. Finally, an additional morphometric analysis was performed on the lordotic juveniles (1 day post-tagging), which were classified on those maintaining their external lordotic phenotype (sevL, severly lordotic) up to the end of the on-growing period, and juveniles that presented a recovered normal phenotype at 434 days post-tagging. The study also included 33 randomly selected normal juveniles.
In all analyses, fourteen landmark measurements were taken on the digital photographs of the examined fish using tpsDig2 software27(Fig. 2). In order to adjust the individuals for centroid size and remove from the landmark configurations any effect irrelevant to shape, a generalized least square method was applied (coordGen6H software package28). TpsRelw software29 was used to calculate the weight matrix (partial warps, with uniform and non-uniform components of shape variation). Canonical variate analysis on the weight matrix was applied to test the effect of lordosis on body shape. Finally, to visualize the variation between the groups, spline diagrams were obtained after the regression of shape components on the canonical scores (tpsRegr software30).
Validation of the phenotypic categorization by means of external morphology
Examination of the external morphology of the 100 juveniles sampled at 1 day post-tagging, revealed that 78 fish were normal (N) and 17 lordotic (L), whereas 5 fish presented an uncertain phenotype (Un) (Fig. 3). The following radiographic examination of the specimens showed that of the 78 N fish, the 74 had a completely normal vertebral column (N/n, Fig. 3), whereas the remaining 4 fish had a lordosis of light severity (N/l-L) (Fig. 3). All fish with lordotic external morphology (L group) presented a severely lordotic vertebral column (L/s-L, Fig. 3). Finally, the five specimens with uncertain external morphology, presented lordosis of light severity (Un/l-L, Fig. 3). In overall, external phenotypic scoring was verified by the radiographic examination in the 95% of normal and 100% of lordotic fish. The applied method of external phenotypic scoring failed to detect the four fish with light internal lordosis, which were categorized as of normal external phenotype (N/l-L).
Evolution of haemal lordosis frequency during the on-growing period
At tagging, phenotypic categorization of the 1700 fish by means of external morphology, showed that the 10.5% of them were lordotic and the 87.0% normal, whereas the 2.5% of the fish were categorized as of uncertain morphology. During the examined on-growing period, the frequency of lordotic fish substantially decreased to 5.8% at 77 dpt (days post-tagging) and 5.2% at 434 dpt (Fig. 4A). The frequency of normal fish increased to 90.2% at 77 dpt and 93.4% at 434 dpt (Fig. 4B).
Changes in the prevalence of normal and deformed fish could not be attributed to a comparatively higher mortality rate of the latter. Results demonstrated that no significant differences existed in the mortality rate between normal and lordotic fish (p > 0.05, G-test). Cumulative mortality rate was 18.7% for normal and 21.8% for lordotic fish (as they were phenotypically scored at tagging) (Fig. 5).
Recovery of haemal lordosis during the on-growing period (1–434 dpt, days-post-tagging)
Monitoring of the external morphology of each abnormal specimen during fish growth, revealed that 20.7% of the fish with lordotic morphology turned into normal phenotype at 77 days post-tagging (dpt). By the end of the on-growing period (434 dpt), 43.6% of the initially lordotic fish presented a completely normal external phenotype (Fig. 6A). At the same time, in 9.3% of the cases, the initially abnormal morphology was turned to uncertain (Fig. 6A). A similar recovery trend was observed in the case of fish with an uncertain external phenotype at 1 dpt, where a substantially higher recovery rate was detected (67.6% at 77dpt, 73.5% at 434 dpt, Fig. 6B).
Recovery of haemal lordosis was verified by the following radiographic examination of the fish at the end of the experiment (434 days post-tagging, dpt). Among fish with recovered external morphology, 26.8% presented a complete recovery of the vertebral column, without any abnormality of individual vertebrae (14.6%, Figs 7A and 8A), or with minor abnormalities of individual centra (12.2%, Figs 7A and 8B). The 48.8% of the fish with recovered external morphology presented a light lordosis (Figs 7A and 8C), whereas 24.4% of them showed a counterbalancing kyphosis anteriorly to lordosis (Figs 7A and 8D). The majority (72.7%) of lordotic specimens which turned to uncertain external appearance at 434 dpt, presented a severe lordosis (Figs 7A and 8E). The partial recovery of haemal lordosis, was also radiographically verified in the case of fish with initially uncertain phenotype (Fig. 7B). Finally, severe vertebral lordosis was present on the radiographies of all the specimens with lordotic external morphology at 434 dpt (Figs 7A and 8F).
Effect of lordosis on the body shape of seabream at the beginning and the end of on-growing period
Canonical variate analysis revealed that lordosis significantly affected the body shape of seabream at both 1 and 434 days post-tagging (Wilks λ = 0.262 and 0.261 respectively, p < 0.001, Fig. 9). In both analyses, the first canonical axis (CV1) explained the majority of phenotypic variance (90.0–93.0%), discriminating normal from lordotic fish. Fish of uncertain phenotype were distributed between normal and lordotic groups (Fig. 9). In the first sampling period, squared Mahalanobis distances were significantly different between all the examined groups (Fig. 9A), whereas in the final sampling only normal fish were significantly different from the lordotic and uncertain groups (Fig. 9B). Spline diagrams demonstrated that abnormal fish were characterized by a ventral transposition of the anterior base of the anal fin (landmark 7) and by a dorsal-anterior shift of the caudal peduncle (landmarks 2–6, Fig. 9).
Morphometric analysis of the recovery of haemal lordosis
Following radiographic examination of fish at the end of the on-growing period (Fig. 7A), we examined whether fish with a recovered normal phenotype differ in body shape from those that were normal since the beginning of the on-growing period (1 day post-tagging, dpt). First axis (CV1) of the canonical variate analysis explained 85.8% of the total variance and discriminated fish with a normal from those with a lordotic phenotype (Wilks λ = 0.155, p < 0.001, Fig. 10). Squared Mahalanobis distances were significant only between lordotic and normal fish (Fig. 10). Distance between fish with a recovered normal phenotype and fish that were normal since the beginning of the on-growing period (1 dpt), was not significant (Fig. 10).
Lordotic juveniles (1 day post-tagging, dpt) were classified according to their morphological evolution into juveniles that maintained the external lordotic phenotype (sevL) up to the end of the on-growing period (434 dpt), and juveniles that presented a recovered normal phenotype at 434 dpt (N-Rec#, Fig. 11). First axis (CV1) of the canonical variate analysis explained 76.3% of the total variance and discriminated fish with a normal phenotype from lordotic sevL and N-Rec# groups (Wilks λ = 0. 420, p < 0.001, Fig. 11). Squared Mahalanobis distances were significant between normal and lordotic fish, independently of the morphological evolution of the latter (Fig. 11). Interestingly, significant body shape differences were detected between lordotic sevL and N-Rec# juveniles (squared Mahalanobis distances, Fig. 11). Spline diagrams demonstrated that shape differences between sevL and N-Rec# lordotic juveniles mainly concerned the severity of the deviation from normal fish (Fig. 11). Use of functions from the canonical variate analysis for the classification of individuals resulted to a successful reclassification of the 94%, 89% and 73% of normal, sevL and N-Rec# fish respectively (Table 3).
This study shows that external morphology of lordotic seabream recovers during the on-growing period. The observed recovery rate depended on the initial abnormality severity and ranged from 43.6% in case of fish with severe lordosis, to 73.5% in case of fish with light severity of lordosis. Interestingly, radiographic examination of fish showed that the recovery of lordotic external morphology was accompanied by a partial to complete recovery of the vertebral column. To our knowledge this is the first study documenting the recovery of severe axis abnormalities in fish. In a similar study, Witten et al.31 demonstrated a remodeling of fused vertebral centra into one non-deformed centrum, which however continued presenting multiple haemal and neural processes.
Decrease of abnormalities prevalence during fish growth is not rare. It is attributed to an increased mortality rate of the abnormal individuals3,10,15,32, or in case of gill-cover abnormalities of light intensity, to the recovery of abnormal phenotype16,17. In the present study, the decreased incidence of lordosis during the growth of Gilthead seabream can not be attributed to a lethal effect of the deformity, since differences in the mortality rate between normal and lordotic fish were not significant. Moreover, the morphological monitoring of pit-tagged individuals during on-growing period revealed that the decrease in the incidence of abnormality was a result of recovery of the lordotic phenotype. In support of this conclusion, at the end of the on-growing period, geometric morphometric analysis showed that there were no significant differences in body shape between fish with a recovered phenotype and those with a normal phenotype since the beginning of the study (Fig. 10). Finally, the recovery of lordosis was also supported by the radiographic examination of fish at the end of the on-growing period (Fig. 7).
In species which are marketed as a whole, establishment of a link between external phenotype and skeleton is a significant goal for the quality assessment of reared fish, since consumers decisions are made on the external morphology26,33. Alterations of body-shape by the presence of haemal lordosis have a continuous range, depending on the angle of the vertebral column at the affected area, and by the number of abnormal vertebrae18,23. With respect to the external morphology, haemal lordosis presents a clear phenotype with a shorter and dorsally shifted caudal penducle, as well as a ventral shift of the posterior abdominal area18,34. In the present study, dorsal shift of the caudal peduncle was used to discriminate lordotic from normal individuals. This method of morphological categorization was successfully validated by radiographic analysis of the juvenile sample taken at the beginning of the study, with 100% of the fish with a lordotic external morphology presenting lordosis of the vertebral column. Applied methodology only failed to correctly categorize 5% (4 out of 78) of the fish with a normal external morphology and a lordotic vertebral column of very light severity (Fig. 3). Significant differences in body shape between juveniles with a normal, lordotic or uncertain external phenotype furthermore supported the initial categorization of the fish (Fig. 9).
Finfish skeleton is subject to continuous resorption, remodeling and reshaping. These processes are essential for development, growth, repair and adaptation of skeleton to mechanical loads35. Haemal lordosis is known to result from the excess mechanical loads of muscles on the vertebral column, during the swimming of the juveniles in hatchery tanks with relatively high water-current speed21,22,23. Since lordotic vertebrae are characterized by increased bone volume, flattened dorsal zygapophyses and extra lateral ridges, lordosis has been suggested as an adaptive response of the vertebral column to the new regime of increased loads36. In the present study, lordosis recovery is shown to be the consequence of two different processes; partial to complete repair of the vertebral column (in 75.6% of the fish with a recovered external morphology), and development of counteracting kyphosis-like bending of the vertebral column (Fig. 8D, in the 24.4% of the fish with a recovered external morphology). Both processes might be triggered by fish transfer to sea cages, an environment of -comparatively to the tanks- less intense water current velocities. Under this hypothesis, observed lordosis-recovery could be a result of the adaptation of growing vertebrae to the new swimming environment.
Despite shape alterations of the deformed vertebrae, histopathological organization of lordotic vertebral centra involves the presence of a fibrous cartilage replacing cancellous and compact bone, enlargement of trabecular spaces and a reduction of the notochordal lumen37. Histological processes that took place during the recovery of lordotic vertebrae in the present paper remain unknown. In the only known similar study on the recovery of vertebral fusion in Atlantic salmon, recovered vertebral bodies acquired the typical normal radiographic appearance, following an histological process involving remodeling of notochordal and cartilage tissues and reshaping of the fused vertebral centra31.
In marine finfish aquaculture, phenotypic quality of the juveniles at the end of the hatchery phase (1–10 g mean weight) has been widely suggested and used as a precise predictor of the phenotypic quality at the end of the on-growing period (>300 g mean weight)1,34,38. Our results clearly suggest that quality control at the end of the hatchery phase has to take into account the recovery potential of haemal lordosis. Interestingly, as it was shown by body shape analysis (Fig. 11), identification of lordotic juveniles which are expected to recover during on-growing is possible through geometric morphometrics. In future, the automatic sort-out of lordotic juveniles with low recovery potential could be possible through the incorporation of computer-assisted systems in the quality control of reared fish.
All data generated and analysed during this study are included in this published article (and its Supplementary Information files).
Boglione, C. et al. Skeletal anomalies in reared European fish larvae and juveniles. Part 2: main typologies, occurrences and causative factors. Rev Aquacult 5, S121–S167 (2013).
Boglione, C. et al. Skeletal anomalies in dusky grouper Epinephelus marginatus (Lowe 1834) juveniles reared with different methodologies and larval densities. Aquaculture 291, 48–60 (2009).
Loizides, M., Georgiou, A. N., Somarakis, S., Witten, P. E. & Koumoundouros, G. A new type of lordosis and vertebral body compression in Gilthead sea bream, Sparus aurata L.: aetiology, anatomy and consequences for survival. J Fish Dis 37, 949–957 (2014).
de Azevedo, A. M. et al. New insight on vertebral anomalies in cultured Senegalese sole (Solea senegalensis, Kaup) at early stages of development. J Fish Dis 40, 987–1000 (2017).
Fragkoulis, S. et al. Saddleback syndrome in European sea bass Dicentrarchus labrax (Linnaeus, 1758): anatomy, ontogeny and correlation with lateral-line, anal and pelvic fin abnormalities. J Fish Dis 40, 83–95 (2017).
Thuong, N. P. et al. Ontogenesis of opercular deformities in gilthead sea bream Sparus aurata: a histological description. J Fish Biol 91, 1419–1434 (2017).
Fernández, I. et al. Larval performance and skeletal deformities in farmed gilthead sea bream (Sparus aurata) fed with graded levels of Vitamin A enriched rotifers (Brachionus plicatilis). Aquaculture 283, 102–115 (2008).
Georga, I. et al. Effect of vitamin A on the skeletal morphogenesis of European sea bass, Dicentrarchus labrax (Linnaeus, 1758). Aquac Res 42, 684–692 (2011).
Cobcroft, J. M., Shu-Chien, A. C., Kuah, M. K., Jaya-Ram, A. & Battaglene, S. C. The effects of tank colour, live food enrichment and greenwater on the early onset of jaw malformation in striped trumpeter larvae. Aquaculture 356-357, 61–72 (2012).
Georgakopoulou, E., Katharios, P., Divanach, P. & Koumoundouros, G. Effect of temperature on the development of skeletal deformities in Gilthead seabream (Sparus aurata Linnaeus, 1758). Aquaculture 308, 13–19 (2010).
Izquierdo, M. S., Socorro, J. & Roo, J. Studies on the appearance of skeletal anomalies in red porgy: effect of culture intensiveness, feeding habits and nutritional quality of live preys. J Appl Ichthyol 26, 320–326 (2010).
Bardon, A. et al. What is the heritable component of spinal deformities in the European sea bass (Dicentrarchus labrax)? Aquaculture 294, 194–201 (2009).
Negrín-Báez, D., Navarro, A., Afonso, J. M., Ginés, R. & Zamorano, M. J. Detection of QTL associated with three skeletal deformities in gilthead seabream (Sparus aurata L.): Lordosis, vertebral fusion and jaw abnormality. Aquaculture 448, 123–127 (2015).
Fragkoulis, S., Batargias, C., Kolios, P. & Koumoundouros, G. Genetic parameters of the upper-jaw abnormalities in Gilthead seabream Sparus aurata. Aquaculture 497, 226–233 (2018).
Koumoundouros, G., Maingot, E., Divanach, P. & Kentouri, M. Kyphosis in reared sea bass (Dicentrarchus labrax L.): ontogeny and effects on mortality. Aquaculture 209, 49–58 (2002).
Beraldo, P. & Canavese, B. Recovery of opercular anomalies in gilthead sea bream, Sparus aurata L.: morphological and morphometric analysis. J Fish Dis 34, 21–30 (2011).
Amoroso, G., Cobcroft, J. M., Adams, M. B., Ventura, T. & Carter, C. G. Concurrence of lower jaw skeletal anomalies in triploid Atlantic salmon (Salmo salar L.) and the effect on growth in freshwater. J Fish Dish 39, 1509–1521 (2016).
Sfakianakis, D. G., Georgakopoulou, E., Kentouri, M. & Koulmoundouros, G. Geometric quantification of lordosis effects on body shape in European sea bass, Dicentrarchus labrax (Linnaeus, 1758). Aquaculture 256, 27–33 (2006).
Kitajima, C., Tsukashima, Y., Fujita, S., Watanabe, T. & Yone, Y. Relationship between uninflated swim bladders and lordotic deformity in hatchery-reared red sea bream Pagrus major. Nippon Suisan Gakk 47, 1289–1294 (1981).
Chatain, B. Abnormal swimbladder development and lordosis in sea bass (Dicentrarchus labrax) and sea bream (Sparus auratus). Aquaculture 119, 371–379 (1994).
Divanach, P., Papandroulakis, N., Anastasiadis, P., Koumoundouros, G. & Kentouri, M. Effect of water currents on the development of skeletal deformities in sea bass (Dicentrarchus labrax L.) with functional swimbladder during postlarval and nursery phase. Aquaculture 156, 145–155 (1997).
Kihara, M., Ogata, S., Kawano, N., Kubota, I. & Yamaguchi, R. Lordosis induction in juvenile red sea bream, Pagrus major, by high swimming activity. Aquaculture 212, 149–158 (2002).
Sfakianakis, D. G. et al. Environmental determinants of haemal lordosis in European sea bass, Dicentrarchus labrax (Linnaeus, 1758). Aquaculture 254, 54–64 (2006).
Mazurais, D. et al. Optimal levels of dietary vitamin A for reduced deformity incidence during development of European sea bass larvae (Dicentrarchus labrax) depend on malformation type. Aquaculture 294, 262–270 (2009).
Sokal, R. R. & Rolhf, F. J. Biometry. The principles and practice of statistics in biological research. 2nd edn, (W.H. Freeman and Company, 1981).
Fragkoulis, S. et al. Scaling of body-shape quality in reared gilthead seabream Sparus aurata L. Consumer preference assessment, wild standard and variability in reared phenotype. Aquac Res 48, 2402–2410 (2017).
tpsDig2, Digitize Landmarks and Outlines v. 2.12 (Department of Ecology and Evolution, State University of New York, Stony Brook, NY, USA 2010).
IMP, Coordinate Generating Utility, CoordGen6h (2001).
TpsRelw, Relative Warps Analysis v. 1.49 (Department of Ecology and Evolution, State University of New York, Stony Brook, NY, USA, 2010).
TpsRegr, Shape Regression v. 1.38 (Department of Ecology and Evolution, State University of New York, Stony Brook, NY, USA, 2011).
Witten, P. E., Obach, A., Huysseune, A. & Baeverfjord, G. Vertebrae fusion in Atlantic salmon (Salmo salar): Development, aggravation and pathways of containment. Aquaculture 258, 164–172 (2006).
Koumoundouros, G., Oran, G., Divanach, P., Stefanakis, S. & Kentouri, M. The opercular complex deformity in intensive gilthead sea bream (Sparus aurata L.) larviculture. Moment of apparition and description. Aquaculture 156, 165–177 (1997).
Losada, A. P. et al. Skeletal malformations in Senegalese sole (Solea senegalensis Kaup, 1858): Gross morphology and radiographic correlation. J Appl Ichthyol 30, (804–808 (2014).
Koumoundouros, G. Morpho-anatomical abnormalities in Mediterranean marine aquaculture. 125–148 (Transworld Research Network, 2010).
Witten, P. E. & Huysseune, A. A comparative view on mechanisms and functions of skeletal remodelling in teleost fish, with special emphasis on osteoclasts and their function. Biol Rev Camb Philos 84, 315–346 (2009).
Kranenbarg, S., Waarsing, J. H., Muller, M., Weinans, H. & van Leeuwen, J. L. Lordotic vertebrae in sea bass (Dicentrarchus labrax L.) are adapted to increased loads. J Biomech 38, 1239–1246 (2005).
Ortiz-Delgado, J. B., Fernández, I., Sarasquete, C. & Gisbert, E. Normal and histopathological organization of the opercular bone and vertebrae in gilthead sea bream sparus aurata. Aquat Biol 21, 67–84 (2014).
Koumoundouros, G. et al. In Success Factors for Fish Larval Production (eds Luís Conceição & Amos Tandler) 435–465 (Wiley Blackwell, 2018).
Τhis study has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 727610 (PerformFISH). This output reflects only the author’s view and the European Union cannot be held responsible for any use that may be made of the information contained therein. We would like to thank E. Karagiannakidis (MD radiologist) and M. Bastouna for their contribution in fish radiography, as well as K. Tzokas, A. Glaropoulos, P. Koutsopodiotis and G. Moukas for their contribution in fish husbandry and sampling.
The authors declare no competing interests.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.