Genome-Wide Scan For Common Variants Associated With Intramuscular Fat And Moisture Content In Rainbow Trout

DOI: https://doi.org/10.21203/rs.3.rs-18231/v1

Abstract

Genetic improvement of fillet quality attributes is a priority of the aquaculture industry. Muscle composition impacts quality attributes such as flavor, appearance, texture, and juiciness. Fat and moisture make up about 80% of the tissue weight. The genetic architecture underlying the fat and moisture content of the muscle are still to be fully explored in fish. A 50K gene transcribed SNP chip was used for genotyping 789 fish with available phenotypic data for fat and moisture content. Genotyped fish were obtained from two consecutive generations produced in the NCCCWA growth-selective breeding program. Estimates of SNP effects from weighted single-step GBLUP (WssGBLUP) were used to perform genome-wide association (GWA) analysis to identify quantitative trait loci (QTL) associated with the studied traits. Using genomic sliding windows of 50 adjacent SNPs, 137 and 178 SNPs were identified as associated with fat and moisture content, respectively. Chromosomes 19 and 29 harbored the highest number of SNPs explaining at least 2% of the genetic variation in fat and moisture content. A total of 61 common SNPs on chromosomes 19 and 29 affected the aforementioned traits; this association suggests common mechanisms underlying intramuscular fat and moisture content. Additionally, based on single-marker GWA analyses, 8 and 24 SNPs were identified in association with fat and moisture content, respectively. SNP-harboring genes were primarily involved in lipid metabolism, cytoskeleton remodeling, and protein turnover. This work provides putative SNP markers that could be prioritized and used for genomic selection in breeding programs.

Introduction

Fillet quality traits have economic importance to the aquaculture industry [1]. For profitable aquaculture production, there is a need for fish fillets with optimum nutritional attributes and consistent organoleptic qualities. Consumer attitude towards fish is influenced by fillet quality, including intramuscular fat and moisture content[2]. Fish provide a valuable source of long-chain omega 3 and other essential fatty acids for human diets, which can lower the risk of several chronic diseases [3]. Moreover, the quantity and quality of intramuscular lipid impact fillet quality attributes, particularly juiciness, flavor, color, texture, and shelf-life [4].

Previous studies showed a high correlation between fat, moisture, and protein content, which together account for 99% of the muscle weight [5]. In mammals, the intramuscular fat content exhibited a significant negative correlation with moisture content [6, 7]. A simultaneous lower protein and moisture content, along with increased fat content of the porcine meat was reported [7]. In fish, previous studies showed that muscle mobilizes protein or fat, replenishing them with water depending on the energetic demands associated with various physiological conditions [8–10]. Muscle degradation, accompanying sexual maturation in female rainbow trout, caused a significant decline in muscle protein content associated with a concomitant increase in moisture content [8]. Meanwhile, fat content was not affected by spawning in rainbow trout, suggesting that intramuscular fat is not required to support gonadal maturation. Additionally, muscle protein catabolism was implied as a source of energy during spawning, generating voids in the muscle that were replaced by water [8]. However, more recent studies in our laboratory revealed that gravid fish maintained on a high plane of nutrition showed reduced intramuscular fat and increased abundance of transcripts involved in fatty acid degradation compared to control sterile fish. These results suggested that female trout approaching spawning mobilize intramuscular fat to fuel gonadal maturation. Spawning fish exhibiting a decrease in muscle fat content showed a concurrent increase in moisture, shear force, and protein content [10]. Many changes in muscle growth and quality attributes occur 2–3 months pre-spawning. Control of the muscle fat content in fish can enhance muscle quality attributes and hence aquaculture industry profitability.

The fat content of the tissue is not usually measured directly on breeding candidates, which makes it difficult for the aquaculture industry to improve it. The aquaculture industry controls the fat content of fillets by adjusting lipid content in the ration [4]. Rearing triploid fish has been adopted in rainbow trout to avoid loss of fillet quality associated with fat mobilization and protein catabolism during sexual maturation [9]. Selective breeding can be used to enhance phenotypic traits of interest. One aim in the breeding programs is to convert feed into muscle rather than accumulating an excess of visceral fat. Also, accumulating excessive lipids in the muscle makes the fillet processing difficult and reduces the fillet firmness [1]. A two-generation program of selection on muscle fat content was initiated in rainbow trout to produce lean and fat lines where the fat percentage increased by ~ 15 to 31% in the fat line depending on the diet [11]. These lines were used as a model to study the effect of muscle fat content on fillet quality [11]. Selection on fat content enhanced fillet color and texture [11], feed conversion ratio (FCR), and protein-retention efficiency [12]. Separately, five generations of family-based selection on body weight of rainbow trout were performed at the USDA National Center of Cool and Cold Water Aquaculture (NCCCWA) [13]. In the fish population of year class (YC) 2010, fat content showed a moderate correlation with whole body weight (regression coefficient (R2) value of 0.50) [14]. Therefore, selection for bodyweight yielded heavier fish with more fat in the muscle. Similarly, gilthead seabream exhibited a 0.1% increase in muscular fat content concomitant with a 0.08% decline in moisture content per increment of ten grams in weight [15]. Muscle fat and moisture content showed moderate heritability in fish, including rainbow trout, implying the existence of genetic variance in a rainbow trout population selected for an enhanced rate of growth [16], thus making genetic responses to selection possible. However, the genetic architecture of fat and moisture content has not been delineated in fish. Understanding the genetic basis of the phenotypic traits in question and development of fish strains of improved genetic gain will enhance efficiency of breeding programs, aquaculture industry profitability, and consumer satisfaction.

Genome-wide association (GWA) studies help to identify variants responsible for phenotypic variations, which can be prioritized in breeding programs that rely on genomic information for making selection decisions. A few GWA studies have been conducted on aquaculture species to identify quantitative genomic loci (QTL) responsible for the genetic variability in body weight [13], fillet quality [13, 17], and disease resistance [18]. Two GWA studies were performed in Atlantic salmon [1] and common carp [19] to identify QTL associated with muscle fat content. In Atlantic salmon, four significant SNPs affecting muscle fat content were identified on chromosomes 9 and 10, based on a 5,650-SNP panel [1]. In common carp, a high-density, 250K SNP array revealed eight SNPs related to muscle fat content; however, none of the SNPs surpassed the genome-wide significance level [19]. The two studies did not identify QTL explaining a large proportion of the genetic variance in fat content for fish. To the best of our knowledge, no GWA studies have been performed in rainbow trout to identify SNP markers associated with genetic variance for fat and moisture content.

A 50K transcribed SNP-chip, suitable for GWA analyses, has been recently developed in our laboratory. The array has been used to identify large-effect QTL responsible for genetic variance in fillet yield, firmness, protein content, and body weight gain [17, 20]. The chip outperformed a 57K genomic SNP panel [13] in identifying major QTL associated with genetic variability in fillet yield and bodyweight gain in fish populations of two consecutive generations of the USDA growth-selection line. The current study aimed to identify QTL affecting additive genetic variance in fillet fat and moisture content for the same rainbow trout population.

Materials And Methods

Fish population, tissue sampling, and phenotypic traits

The fish population used in the current GWA analyses was previously described in detail [21]. Briefly, a selective breeding program was established at NCCCWA in 2002 by intercrossing seven domesticated strains of rainbow trout; this fish population has gone through five generations of selection for improved growth performance [22]. Full-sib families collected from third- and fourth-generation fish (Year-class; YC 2010 and 2012) were used for the current GWA analyses. Phenotypic data for muscle fat and moisture content were obtained from 789 fish representing 197 families produced from two generations (YC 2010 and 2012). Single-sire × single-dam matings occurred over 6 weeks to produce full-sib families. Individuals from each family were reared together in a 200-L tank in order to keep the pedigree information. Tagging fish, at ~ 5-months post-hatch with a passive integrated transponder (Avid Identification Systems Inc., Norco, CA) allowed husbandman to rear different fish families together in 800-L communal tanks (~ 13 °C) until ~ 13 months post-hatch. A commercial fishmeal-based diet was provided, and the feeding schedule was previously described [23]. Fish were purged for 5 days before harvest to facilitate viscera removal.

Whole-body weight (WBW) was measured, and fish were sorted within family accordingly. For sampling, the 2nd or 3rd fish from each sorted family was selected in order to keep the distribution of WBW around the median of each family. For each YC, fish were randomly assigned to one of five harvest groups allowing a single fish per family per harvest group (~ 100 fish each). The harvest groups were sampled over five weeks (one group/week). Fish groups from the two-year classes were harvested at different ages. Fish produced from YC 2010 were harvested between 410- and 437-days post-hatch (mean body weight = 985 g; SD = 239 g), whereas, fish produced from YC 2012 were harvested between 446- and 481-days after hatching (mean body weight = 1,803 g; SD = 305 g). Fish were euthanized with tricaine methanesulfonate (Tricaine-S, Western Chemical, Ferndale, WA), harvested, and eviscerated. Head-on gutted carcasses were shipped on ice to the West Virginia University Muscle Foods Processing Laboratory (Morgantown, WV) where carcasses were manually processed into skinless fillets [10].

Proximate analyses, including crude lipid and moisture content, were previously described [24]. In brief, crude lipid and moisture analyses were performed using AOAC-approved methods (AOAC 2000). Crude lipid content was determined using Soxhlet extraction with petroleum ether. Moisture content was assessed by weighing the sample before and after drying at 110 °C for 18 hrs. When muscle fat and moisture content were regressed on body weight, regression coefficient (R2) values of 0.23 and 0.38 were observed, respectively. The pedigree-based heritability h2 (h2ped) for fat and moisture content was estimated according to Zaitlen et al. [25].

SNP genotyping and quality control

A 50K, transcribed gene SNP-chip was recently developed and used in identifying genomic loci responsible for additive genetic variance in fillet yield [17]. SNPs used to build the SNP chip were identified and reported in our previous study [21]. Briefly, the chip included ~ 21K SNPs showing potential allelic imbalances with growth, muscle quality, and susceptibility/resistance to Bacterial Cold Water Disease (BCWD) [17, 21]. About 5K nonsynonymous SNPs were included in the chip. Additional SNPs were added to the chip to have a minimum of 2 SNPs per each SNP-harboring gene. The array contains 50,006 SNPs.

In total, 1,728 rainbow trout fish were used for genotyping and quality assessment of the SNP chip. Genotyped fish were collected from the USDA/NCCCWA growth- and BCWD-selection lines [17]. Samples and genotypes had undergone quality control checks. The Affymetrix SNPolisher software was used for QC assessment and filtration of samples/genotypes at the default settings [26]. Genotyped samples were filtered using a call rate of 0.97 and Dish QC (DQC) threshold of 0.82. In the current study, 789 genotyped fish, with available phenotypic data for intramuscular fat and moisture content, were used for GWA analyses.

Fifty-SNP window GWA analysis

Estimates of SNP effects from weighted single-step GBLUP (WssGBLUP) were used to conduct the current GWA analysis as we previously described [17]. The WssGBLUP combines phenotypes, genotypes, and pedigree information into a single evaluation, where SNPs used to construct the genomic relationship matrix are weighted according to the proportion of additive variance they explain. The following a mixed model was used:

y = Xb + Z1a + Z2w + e

where y is the vector of phenotypes, b is the vector of fixed effects (harvest group and hatch-year), a is the vector of random animal effect, w is the vector of random family effect, and e is the vector of residual effects. X, Z1, and Z2 are incidence matrices for the effects contained in b, a, and w, respectively. Whereas the family and residual random effects were considered uncorrelated, the animal effect was correlated. The covariance structure for the animal effect was given by

where H is a matrix that combines pedigree- and genomic-based relationships[27] and

is the additive variance.

The variance components were estimated using AIREMLF90[28]. The inbreeding coefficient was calculated by INBUPGF90 [17, 29] using pedigree data representing 63,808 fish produced from five consecutive generations by the NCCCWA breeding program. Genomic data were edited using PREGSF90[28], and samples or SNP were kept according to the following parameters: minor allele frequency (MAF) > 0.05, call rate > 0.90, and Hardy-Weinberg equilibrium (HWE) < 0.15. Out of 50,006 SNPs, 35,322 SNPs (70.6%) passed QC and were used for the WssGBLUP analysis.

Two iterations were used in the current WssGBLUP analysis. All SNPs passing QC were assigned, by default, the same weight in the first iteration (i.e., weight = 1.0). In the second iteration, SNP weights were determined according to SNP effects (û) calculated in the first iteration as asû22p(1 - p), where p is the allele frequency in the current genotyped population. In brief, three steps were conducted for each iteration: 1) weights were assigned to the SNPs; 2) genomic estimated breeding values (GEBV) were calculated, based on H− 1 [29]; and 3) SNP effects and updated weights were computed by POSTGSF90 [29] using genomic sliding windows of 50 adjacent SNPs. This window size (n = 50 SNPs) was adopted instead of physical size (e.g. specific number of nucleotides) because SNPs were not evenly distributed over the whole genome. The qqman package [30] in R was used to generate Manhattan plots showing the proportion of additive variance explained by the 50 SNP windows.

Single marker GWA analysis

PLINK [31] was used to filter the genomic data before performing a single marker association analysis. The filtering criteria included, MAF > 0.05 and HWE < 0.001. The command, --assoc, was used to retrieve R-squared values (R2) of association between the quantitative traits and genotypes. A generalized test score in ONETOOL [32] was used to perform family-based association analysis. This test allows for multiple covariates and accounts for family structure using a kinship matrix. Covariates were incorporated in the linear model to account for fixed effects (harvest group and hatch-year) and population structure. The first two principal components (PCs) were included in the model to account for population structure. P-values with Bonferroni correction were used to account for multiple testing. Manhattan plots showing single SNP markers associated with variations in intramuscular fat and moisture content were generated using qqman package [30].

Gene annotation

To annotate the SNP-harboring genes, a bed file containing SNPs associated with the quality trait of interest was intersected with the rainbow trout genome gff/gtf file using Bedtools [17, 33]. SNPs were classified as genic or intergenic according to their physical position relative to the body of the gene. Genic SNPs exist in coding DNA sequence (CDS), introns, or untranslated regions (5’UTR and 3’UTR). Intergenic SNPs are defined as SNPs located in the region between genes.

Results And Discussion

Muscle fat and moisture contents are interrelated attributes that affect the organoleptic quality and nutritional value of muscle foods [19]. In fish, high-fat content may influence fillet processing and reduce the firmness leading to fillet downgrading [1]; moreover, it significantly impacts texture, juiciness, and flavor. In mammals, increased marbling scores are positively related to beef tenderness, accounting for ~ 9% of the shear force variation [34]. The inability to retain moisture during postmortem storage, in both fish and mammals, is associated with a high drip loss and, in turn, reduces the industry profitability by influencing processing yield and palatability [35, 36]. In the pork industry, drip loss results in up to 10% product losses affecting profitability at wholesale and retail levels [35]. Increased knowledge of the genetic basis of muscle quality traits will facilitate to further the commercial breeding in salmonids. GWA studies are powerful tools to identify genetic variants associated with complex traits [19]. However, no GWA studies were previously conducted to dissect the genetic architecture of fillet fat and moisture contents in rainbow trout. The estimated heritability for fat and moisture content was 0.31 ± 0.07 and 0.42 ± 0.07, respectively, suggesting existence of adequate genetic variability in the NCCCWA fish population to allow genetic improvement through selective breeding. A higher rate of genetic gain is obtained when genomic information is used [37].

In this study, we used genomic windows of 50 SNPs of a 50K SNP chip, to perform GWA analyses, in addition to the single-marker analysis approach, to identify genomic regions associated with the traits. Given that the 50 K, SNP chip contains SNPs of potential association with intramuscular fat content [28; 49], all fish used to build the SNP-chip were excluded from the GWA analysis in the present study.

The fish population used for the current GWA analyses had an average muscle fat content of 9.2 ± 1.91 (%) and moisture content of 69.93 ± 1.75 (%). Variations in fat and moisture content are shown in Figure (1). Previous studies reported a significant correlation between changes in fat and moisture content in fish [10, 15]. Consistently, our data showed a significant negative correlation between fat and moisture content (R = -0.88; p-value = 6.3E-262).

QTL affecting muscle fat and moisture content using WssGBLUP

All 35,322 SNPs (70.6%) that passed QC were used in the WssGBLUP analysis. A complete list of proportions of additive genetic variance for fat content explained by all genomic windows is provided in Table S1. Of them, a total of 137 genomic sliding windows explaining at least 2% (arbitrary value) of the additive genetic variance for fat content are listed in Table S2. Most of the SNP sliding windows (n = 124; ~91%) were located within 62 protein-coding genes. Genomic loci affecting the additive variance for fat content were clustered in 5 chromosomes (1, 4, 5, 19, and 29) (Fig. 2).

Chromosome 19 harbored the highest number (n = 50) and the most significant peaks affecting fat content (up to 5.51%) (Table S2, Fig. 2). Many of the SNPs were located within the CDS of the SNP-harboring genes (n = 58) as well as their 3’UTR (n = 55). In order to understand the biological significance of the QTL associated with fat content, we annotated the SNP-harboring genes and searched their functions in the literature (described below).

Similarly, a complete list of the proportions of additive genetic variance for moisture content explained by all windows identified in this study is provided in Table S3. A total of 178 genomic sliding windows revealing at least 2% of the additive genetic variance for moisture content are listed in Table S4. Most of the SNP sliding windows (n = 165; ~93%) were located within 86 genes coding for proteins. Genomic loci affecting the additive variance for moisture content were clustered on 5 chromosomes (5, 14, 19, 25, and 29) (Fig. 3). Chromosome 29 harbored the highest number (n = 48), whereas the most significant peaks affecting moisture content (up to 4.46%) were identified on chromosome 19 (Table S4, Fig. 3). Many of the SNPs were located within CDS of the SNP-harboring genes (n = 68) as well as their 3’UTR (n = 72).

Common genes affecting muscle fat and moisture content

As shown above, a negative linear relationship has been established between fat and moisture content in this selectively bred rainbow trout population (YC 2010 and YC 2012), suggesting a common mechanism underlying the genetic variation in the two traits. This negative correlation was consistent with other studies in fish and mammals [6, 7, 10, 15]. The current WssGBLUP identified common SNPs affecting the additive genetic variance for fat and moisture content on chromosomes 19 and 29 (Tables S2 & S4). The majority of the common SNPs (n = 47) were located on chromosome 19. Thirty-two SNPs, out of 47, involved in lipid metabolism were identified in 16 protein-coding genes on chromosome 19 (Table 1). Briefly, cathepsin B had a single 3’UTR SNP. Cathepsin B regulates very-low-density lipoprotein (VLDL) secretion and free fatty acid uptake in response to oleic acid exposure in mice [38]. Thioredoxin-related transmembrane protein 1-like (TMX1) had three SNPs. Loss of TMX increases lipid peroxidation in TMX−/− mice, which, in turn, enhances oxidative stress [39]. Guanine nucleotide-binding protein GI/GS/GO gamma-2 subunit (GNG2) had a single 3’UTR SNP. GNG2 expression is positively correlated with adipocyte size [40]. SNPs in genes encoding beta-taxilin and Alpha-L-fucosidase 2 (FUCA2) were covering windows explaining the highest proportion of the additive genetic variation for fat and moisture content. Adipose tissue of obesity susceptible and resistant rats differentially expressed beta-taxilin under a high-fat diet [41]. FUCA2 is a glycolipid processing enzyme [42]. Two SNPs in F-box only protein 30 (FBXO30) and the microtubule-binding protein ensconsin were ranked next to beta-taxilin and FUCA2. An SNP in FBXO30 was located in a genomic region explaining 4.95% of the additive genetic variance for polyunsaturated fatty acids in cattle [43]. Knock-down of microtubule-binding or -associated proteins led to changes in fat accumulation during adipogenesis [44]. Dihydropyrimidinase-related protein 5-like (CRMP5) had a single synonymous SNP. CRMP5 has GO terms belong to lipid metabolic processes [45]. Five SNPs were identified in a gene encoding trifunctional enzyme subunit alpha, mitochondrial (HADHA). This gene is involved in fatty acid beta-oxidation [45]. mRNA decay activator protein ZFP36L1 had three SNPs in the 3’UTR. Knockdown of mammalian ZFP36L1 led to downregulation of ERK activation and inhibition of adipogenesis [46]. A single 3’UTR SNP was identified in ELM2 and SANT domain-containing protein 1 (ELMSAN1). Epigenome-wide association analysis showed DNA methylation changes in ELMSAN1 were associated with body mass index (a key measure of adiposity) [47]. Prostaglandin reductase 2 (PTGR2) had two nonsynonymous SNPs. This enzyme catalyzes reduction of the conjugated α,β-unsaturated double bond of 15-keto-PGE2 in an NADPH-dependent manner, which is a critical step in inhibition of PPARγ-mediated adipocyte differentiation [48]. Spectrin beta chain, erythrocytic (SPTB) gene had two SNPs. The SPTB interacts with phospholipids in natural membranes and has a role in controlling the fluidity of the inner lipid leaflet of the cell membrane [49].

Chromosome 29 had 14 SNPs in genomic windows explaining at least 2% of the additive genetic variance for intramuscular fat and moisture content (Tables S2 & S4). Of them, seven SNPs were involved in lipid metabolism (Table 1). A single SNP was identified in a gene encoding short-chain specific acyl-CoA dehydrogenase, mitochondrial (ACADS). This enzyme has a role in fatty acid beta-oxidation [50]. An intronic SNP was identified in a gene coding for arrestin domain-containing protein. The latter has GO terms belonging to fat pad and skin development and regulates the body mass [51]. Myocyte enhancer factor 2c (MEF2C) had the highest number of SNPs (n = 5) on chromosome 29. MEF2C is a transcription factor involved in skeletal muscle differentiation; however, it has been reported as a constituent of a mechanism that programs gene expression involved in development of brown adipocytes [52]. MEF2A and MEF2D isoforms exhibited in vivo differential expression in mammalian striated muscle and white adipose tissue of insulin-deficient diabetic mice [53]. To our knowledge, the role of MEF2C in white adipose tissue remains uncertain.

In addition, twelve SNPs in genes involved in transmembrane transport and cytoskeleton remodeling were identified in common QTL affecting additive variance for fat and moisture content (Table 2). The majority of these SNPs were identified on chromosome 19 (n = 11). Three synonymous SNPs were identified in a gene encoding intersectin-2 (ITSN2). This protein is necessary for the clathrin-mediated endocytosis and actin cytoskeleton remodeling [54]. Six SNPs were identified in 3 genes involved in vesicle-mediated transport (i.e., exocytosis); dnaJ homolog subfamily C member 5B, visinin-like protein 1, and syntaxin-binding protein 5. The actin cytoskeleton remodeling controls each step of exocytosis [55]. Three SNPs were identified in microtubule-associated protein RP/EB family member 3 (MAPRE3) and centrin-3. MAPRE3 and centrin-3 control the dynamics of the microtubule cytoskeleton [45, 56].

Unique genes affecting the additive genetic variance for moisture

The actin cytoskeleton interacts with the cell membrane to control water transport [57]. In the current study, thirty-six variants in genes (n = 14) involved in cytoskeleton remodeling were identified, affecting the additive variance for moisture content in rainbow trout (Table 3). Briefly, bone morphogenetic protein receptor type-2 (BMPR2) had a single synonymous SNP. BMPR2 is known to interact with the cytoskeleton, and BMPR2 mutant mice exhibited cytoskeletal defects [58]. A single SNP was identified in a gene encoding muscle associated receptor tyrosine kinase (MUSK). Activation of MUSK in myotubes regulates the reorganization of the actin cytoskeleton [59]. Two SNPs were identified in THAP domain containing 1 (THAP1), which has a role in regulation of the mitotic cell cycle [60]. The gene encoding asparaginyl-tRNA synthetase (NARS) had 3 SNPs in windows explaining the highest additive variance (up to 3.46%; Table 3). Mutations in NARS leads to cell cycle arrest in the S phase [61]. The actin cytoskeleton undergoes dramatic changes during the cell cycle [62]. Ten SNPs were identified in three genes coding for cyclin-I (CCNI), cyclin-G1 (CCNG1), and cyclin-G2 (CCNG2). Cyclins function as regulators of the cell cycle and actin cytoskeleton dynamics [63]. The serine/threonine-protein, phosphatase 2A (PP2A), had a 3’UTR SNP. This phosphatase is associated with microtubule stabilization, where it binds and dephosphorylates the microtubule-associated proteins [64]. Annexin A6 (ANXA6) had two synonymous SNPs. ANXA6 contributes to membrane and cytoskeleton organization in a Ca2+-dependent manner [65]. Tubulin beta-4B chain (TUBB4B) had four synonymous SNPs within 1Kb of chromosome 25. TUBB4B is a critical component of microtubules [45]. Five SNPs, clustered in ~ 2Kb, were identified in a gene coding for mid1-interacting protein 1 (MID1IP1) [66]. This protein enhances fatty acid biosynthesis [66] and stabilizes microtubule organization [45]. Two SNPs were identified in a gene encoding tubulin-specific chaperone A (TBCE). TBCE is a tubulin-folding protein required for proper microtubule cytoskeleton organization [67]. Additionally, mutations in TBCE drive muscle atrophy [67]. Proteinase-activated receptor 1 (PAR1) and PAR2 had four SNPs. PAR-mediated, RhoA activation is vital for cytoskeletal reorganization [68].

A strong positive correlation between moisture and protein content has been established in different species. A simultaneous decline in protein and moisture content was previously reported in mammals [7]. Moisture content in rainbow trout exhibited a bidirectional relationship with protein content depending on physiological/metabolic status. For example, a negative correlation between moisture and protein content were previously reported under muscle catabolic conditions associated with full sexual maturation (R2 = 0.994, p < 0.01) [8]; whereas, a positive correlation was reported in female trout, on a high plane of nutrition, that were approaching spawning [10]. This was explained by selective mobilization of either protein during spawning or fat before spawning; in either case, the depleted macromolecule was replaced by water. It is noteworthy that protein content variation of the current study was not statistically significant between the 4 high-ranked families versus 4 low-ranked families (data not shown). The current WssGBLUP analysis indicated that thirteen SNPs in genes involved in protein degradation were involved in the additive genetic variance of moisture content (Table 4). Briefly, E3 ubiquitin-protein ligase RNF170 is an E3 ubiquitin-protein ligase that plays an essential role in the ubiquitination and degradation of inositol 1,4,5-trisphosphate receptor type 1 (ITPR1). The latter controls the calcium release from the endoplasmic reticulum [69], which affects the muscle protein content in rainbow trout [20] and has a profound effect on the regulation of cytoskeleton [70]. Cystatin-1, which possesses a peptidase inhibitor activity, had a single 5’UTR SNP. Thioredoxin-like 1 (TXNL1) had two synonymous SNPs. The knockdown of TXNL1 moderately stabilizes the ubiquitin-protein conjugates suggesting a connection between protein reduction and proteolysis [71]. Pre-mRNA-processing factor 19 (PRPF19) and ubiquitin-conjugating enzyme E2 D2 (UBE2D2) had four SNPs. These ligases catalyze polyubiquitin chain assembly and play a role in proteasomal protein degradation [72, 73]. Nuclear factor NF-kappa-B p105 subunit (NFKB1) had two 3’UTR SNPs. NFKB1 is involved in the negative regulation of cellular protein metabolic process [74] and apoptotic process [75].

In addition to the ubiquitin-protein ligases, SNPs in three genes involved in lysosomal/phagosomal pathways were identified. Ras-related protein rab7 (RAB7A) harbored a 3’UTR SNP. RAB7A is a major regulator of endo-lysosomal maturation/ trafficking and protein targeting to lysosome inducing autophagosome formation [76]. Thus, RAB7A positively regulates the protein catabolic process [77]. V-type proton ATPase subunit B (ATP6V1B2) had a 3’UTR SNP. V-ATPase is responsible for acidifying the intracellular compartments, including lysosomes [78]. The gene encoding the β chain of the adaptor protein-3 (AP-3) complex had a single synonymous SNP. Deletion in AP3B1 perturbs assembly of AP-3 complex and, in turn, trafficking of transmembrane lysosomal proteins [79].

Taken together, most of the common genomic loci affecting the highest proportion of the additive variance were involved in lipid metabolism, suggesting a common mechanism underlying intramuscular fat and moisture content and, partially, explaining the strong negative correlation between the fat and moisture content in this selectively bred rainbow trout population. Unique loci affecting moisture content were primarily involved in cytoskeleton regulations and protein turnover. Inhibition protease activity, such as calpains, reduced degradation of proteins responsible for cell membrane-cytoskeleton attachments and postmortem drip channel formation in muscle. The presence of calcium enhances proteolysis by µ-Calpain, of myofibrillar and other cytoskeletal proteins during postmortem storage [36].

Single marker GWA analyses

To identify single SNP markers associated with variation in fat and moisture content, we analyzed SNPs that passed QC filtration (n = 29,451) using a generalized score test; this test incorporates multiple covariates in the analysis and accounts for family structure using a kinship matrix [32]. In this study, 8 and 24 significant SNPs, surpassing the genome-wide significance level, had a potential impact on the fat and moisture content (Bonferroni-corrected p < 1.69E-06; Figs. 4 & 5 and Tables S5 & S6), respectively.

SNPs associated with the fat content were mainly located on chromosome 5 (n = 7) and SNP-harboring genes have roles in lipid metabolism (Table 5). The list includes 78 kDa glucose-regulated protein (GRP78), spindle and kinetochore associated complex subunit 1 (SKA1), apelin receptor B (APLNR-B), desmoplakin, podocan, and calcium-binding mitochondrial carrier protein SCaMC-1 (SLC25A24). Briefly, two missense mutations were identified in genes coding for GRP78 and SKA1. GRP78 is essential for adipocyte differentiation and a balanced secretion of adipokines. Deletion of GRP78 causes lipoatrophy in mice observed as a dramatic reduction in gonadal and subcutaneous adipose tissue [80]. SKA1 was downregulated in adipose tissues between samples from obese and healthy control children and has been suggested as a candidate biomarker for childhood obesity [81]. Three synonymous mutations were identified in genes encoding APLNR, desmoplakin, and podocan. APLNR knockout mice demonstrated excess fatty acid accumulation in skeletal muscle [82]. Abnormalities in desmoplakin have been associated with changes in lipid metabolism [83]. Podocan belongs to the small leucine-rich proteoglycans (SLRPs) that bind to low-density lipoprotein receptor-related protein (LRP-1) [84]. A 3’UTR SNP was identified in a gene coding for SLC25A24. Mice fed a high-fat diet exhibited increased expression level of SLC25A24; whereas, adipocyte differentiation was suppressed in Slc25a24- knockout [85].

Puromycin-sensitive aminopeptidase (NPEPPS), on chromosome 17, had a nonsynonymous SNP explaining the highest variability in fat content (R2 = 3.2%) (Table 5). NPEPPS impacts different physiological processes, including protein turnover and cell cycle regulation. NPEPPS was upregulated in mitten crabs fed with a linseed oil rich in linoleic acid [86]. However, the effect NPEPPS on lipid metabolism in fish needs further investigations.

SNPs associated with moisture content (n = 24) were associated with protein turnover, calcium metabolism, and cytoskeleton integrity (Table S5). Most of these SNPs (n = 11; ~46%) were located on chromosome 17. An SNP in a gene coding for acylphosphatase 2 ranked at the top of the list (R2 = 7.4%) (Table 5); however, its physiological role is not clear. Eight SNPs associated with moisture content were identified in five genes engaged in protein metabolism (Table 5). These genes are NPEPPS, eukaryotic initiation factor 4A-III (EIF4A3), eukaryotic translation initiation factor 4B (EIF4B), ribosome binding protein 1 (RRPB1), and F-box only protein 46 (FBXO46). Briefly, the aminopeptidase, NPEPPS, was associated with variation in fat and moisture content, suggesting a correlation between moisture and fat content. Five SNPs were identified in two genes encoding EIF4A3 and EIF4B, suggesting a role for the protein translation machinery in determining variation in moisture content. RRPB1 is an ER integral membrane protein implicated in polysome assembly and, therefore, protein synthesis [87]. RRPB1 has been suggested as essential in regulation of UPR signaling molecules and autophagy [88]. Finally, the F-box family SCF-E3 ubiquitin ligase, FBXO46, had a single 3’UTR SNP.

Expression of genes involved in the actin cytoskeleton and cytoskeletal organization was positively correlated with drip loss [89]. In this study, a total of 10 SNPs associated with moisture content were identified in seven genes engaged in cytoskeleton regulation (Table 5). These genes are encoding serum response factor (SRF), kinesin-1 heavy chain (KIF5B), inositol-trisphosphate 3-kinase C (IP3KC), supervillain (SVIL), calcineurin subunit B type 1 (PPP3R1), eIF4A and EIF4B. Briefly, SRF is a master regulator of the actin cytoskeleton [90]. Mutations in KIF5A caused cytoskeletal defects in humans [91]. Two, 3’UTR SNPs were identified in IP3KC and PPP3R1. IP3K and calcium/calcineurin signaling play critical roles in maintaining Ca2+ homeostasis that has a profound effect on the cytoskeleton [70, 92, 93]. Supervillin (SVIL) is one of the first components of the costameric membrane skeleton to assemble during muscle formation. It establishes a high-affinity connection between the membranes and actin cytoskeleton [94]. Translation initiation factors, including eIF4A and EIF4B, associate with the actin cytoskeleton, which affects protein synthesis [95].

In our previous work, we profiled transcriptome expression of fish families (YC 2010) showing contrasting phenotypes in fat content, which revealed only 17 differentially expressed transcripts associated with fat content [14]. About 90% of the genetic variation among individuals comes from SNPs [96], and therefore, identifying SNP markers associated with complex traits is most suitable for genetic evaluation in selection programs. Two previous GWA studies identified a few SNPs responsible for the additive variance for fat content in Atlantic salmon and Common Carp [1, 19]. The current GWA analysis identified a total of 137 SNPs in windows explaining at least 2% of the additive genetic variance for fat content, suggesting a better characterization of the genetic basis underlying variation in fat content. The discrepancies among the different GWA studies might be due to; 1) usage of different algorithms in the GWA studies, 2) variation in population size, 3) substantial difference in the capacity of the SNP arrays, 4) polygenic nature of intramuscular fat content, 5) different thresholds in each study including sliding window size [13].

Compared to our WssGBLUP analysis, the single marker GWA analysis revealed a smaller number of SNP markers associated with variation in intramuscular fat and moisture content. In addition, these two GWA approaches revealed different significant peaks associated with traits of interest. Aguilar et al. (2019) showed that the highest peak based on the p-value was not the same based on the proportion of variance explained, and this is because the latter depends on allele frequency, i.e., high effect but low frequency decreases the variance explained[97]. This result is consistent with other studied traits, such as fillet firmness, protein content [20], and bodyweight gain [98] in rainbow trout. The potential factors associated with observed heterogeneity between the two approaches are different algorithms, thresholds, and windows size used in each approach. The WssGBLUB was more effective than the single marker GWA in examining the genetic architecture of studied traits and identifying common QTL between traits. This method has proven to be optimal for breeding populations given the data structure: phenotyped individuals may not have genotypes, and there is a long history of pedigree recording [97]. Common QTL identified in this study may explain the high negative correlation between fat and moisture content. The recombinational progression of QTL and nearby markers determines the information content of haplotypes [99]. However, SNP-harboring genes identified by the two approaches had similar biological functions and were involved in lipid metabolism, protein turnover, and cytoskeletal remodeling. Routine use of single-SNP and multi-makers for GWA analysis was previously recommended to take advantage of the complete information content of the genotypes [99].

Conclusions

The current GWA analyses identified novel genomic regions associated with additive genetic variance for fat and moisture content in rainbow trout. These genomic loci code for proteins involved in lipid metabolism, actin cytoskeleton remodeling, and protein synthesis/degradation. Compared to a previous GWA work in Atlantic salmon, this work reveals significant QTL affecting fat content, which appears to be a polygenic trait. The top common windows affecting additive genetic variance for fat and moisture content are mainly on chromosome 19. These findings provide a genetic basis for description of the molecular mechanisms underlying fat and moisture content in teleost fish. In addition, this work provides putative markers that could be prioritized when estimating genomic breeding values for fat and moisture content, aiming an increased accuracy.

Abbreviations

BCWD

Bacterial Cold Water Disease

GO

Gene ontology

GWA

Genome-wide association

HWE

Hardy–Weinberg equilibrium

MAF

Minor allele frequency

NCCCWA

USDA National Center of Cool and Cold Water Aquaculture

QC

Quality control

QTL

Quantitative trait loci

SNP

Single nucleotide polymorphism

UTR

Untranslated region

WssGBLUP

Weighted single-step GBLUP

YC

Year class

Declarations

Ethics approval and consent to participate

Institutional Animal Care and Use Committee of the United States Department of Agriculture, National Center for Cool and Cold Water Aquaculture (Leetown, WV) specifically reviewed and approved all husbandry practices used in this study (IACUC protocol #056).

Consent for publication

Not applicable.

Competing interests

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Funding

This study was supported by a competitive grant No. 2014-67015-21602 from the United States Department of Agriculture, National Institute of Food and Agriculture (MS), and by the USDA, Agricultural Research Service CRIS Project 1930-31000-010 “Utilizing Genetics and Physiology for Enhancing Cool and Cold Water Aquaculture Production”.  The content is solely the responsibility of the authors and does not necessarily represent the official views of any of the funding agents.

Authors' Contributions

MS, TL, and BK conceived and designed the experiments. RA-T, MS, TL, and BK performed the experiments. RA-T, AA, DL, BK, and MS analyzed the data. AA and MS wrote the manuscript.

Acknowledgments

The authors acknowledge J. Everson, M. Hostuttler, K. Jenkins, J. Kretzer, J. McGowan, K. Melody, T. Moreland, and D. Payne for technical assistance. The use of trade, firm, or corporation names in this publication is for the information and convenience of the reader.  Such use does not constitute an official endorsement or approval by the USDA or the ARS of any product or service to the exclusion of others that may be suitable.  USDA is an equal opportunity provider and employer.

Availability of data and materials

All datasets generated for this study are included in the manuscript and/or the Additional Files. The genotypes (ped and .map files) and phenotypes are available in our previous publication [20].

References

  1. Sodeland M, Gaarder M, Moen T, Thomassen M, Kjøglum S, Kent M, Lien S: Genome-wide association testing reveals quantitative trait loci for fillet texture and fat content in Atlantic salmon. Aquaculture 2013, 408–409:169–174.
  2. Tumuluru JS, Sokhansanj S, Bandyopadhyay S, Bawa AS: Changes in Moisture, Protein, and Fat Content of Fish and Rice Flour Coextrudates during Single-Screw Extrusion Cooking. Food and Bioprocess Technology 2013, 6:403-415.
  3. Zhao F, Zhuang P, Song C, Shi Z, Zhang L: Amino acid and fatty acid compositions and nutritional quality of muscle in the pomfret,  Pampus punctatissimus. Food Chemistry 2010, 118:224–227  
  4. Manor M, Weber G, Salem M, Yao J, Aussanasuwannakul A, al. e: Effect of sexual maturation and triploidy on chemical composition and fatty acid content of energy stores in female rainbow trout,  Oncorhynchus mykiss. Aquaculture 2012, 364:312–321.
  5. Ang CYW, Young LL, Wilson RJ: Interrelationships of protein, Fat and Moisture Content of broiler Meat. Food Sci 1984, 49:359  
  6. Watanabe G, Motoyama M, Nakajima I, Sasaki K: Relationship between water-holding capacity and intramuscular fat content in Japanese commercial pork loin. Asian-Australas J Anim Sci 2018, 31(6):914-918.
  7. Li YX, Cabling MM, Kang HS, Kim TS, Yeom SC, Sohn YG, Kim SH, Nam KC, Seo KS: Comparison and correlation analysis of different Swine breeds meat quality. Asian-Australas J Anim Sci 2013, 26(7):905-910.
  8. Salem M, Kenney PB, Rexroad CE, Yao J: Molecular characterization of muscle atrophy and proteolysis associated with spawning in rainbow trout. Comp Biochem Physiol Part D Genomics Proteomics 2006, 1(2):227-237.
  9. Paneru B, Ali A, Al-Tobasei R, Kenney B, Salem M: Crosstalk among lncRNAs, microRNAs and mRNAs in the muscle 'degradome' of rainbow trout. Sci Rep 2018, 8(1):8416.
  10. Salem M, Manor ML, Aussanasuwannakul A, Kenney PB, Weber GM, Yao J: Effect of sexual maturation on muscle gene expression of rainbow trout: RNA-Seq approach. Physiol Rep 2013, 1(5):e00120.
  11. Florence L, Mireille C, Jérôme B, Laurent L, Françoise M, Edwige Q: Selection for muscle fat content and triploidy affect flesh  quality in pan-size rainbow trout, Oncorhynchus mykiss  Aquaculture 2015, 448:569-577.
  12. Kause A, Kiessling A, Martin SA, Houlihan D, Ruohonen K: Genetic improvement of feed conversion ratio via indirect selection against lipid deposition in farmed rainbow trout (Oncorhynchus mykiss Walbaum). Br J Nutr 2016, 116(9):1656-1665.
  13. Gonzalez-Pena D, Gao G, Baranski M, Moen T, Cleveland BM, Kenney PB, Vallejo RL, Palti Y, Leeds TD: Genome-Wide Association Study for Identifying Loci that Affect Fillet Yield, Carcass, and Body Weight Traits in Rainbow Trout (Oncorhynchus mykiss). Front Genet 2016, 7:203.
  14. Ali A, Al-Tobasei R, Kenney B, Leeds TD, Salem M: Integrated analysis of lncRNA and mRNA expression in rainbow trout families showing variation in muscle growth and fillet quality traits. Sci Rep 2018, 8(1):12111.
  15. García-Celdrán M, Ramis G, Manchado M, Estévez A, Navarro A: Estimates of heritabilities and genetic correlations of raw flesh quality traits in a reared gilthead sea bream ( Sparus aurata  L.) population sourced from broodstocks along the Spanish coasts. Aquaculture 2015, 446:181-186.
  16. Leeds T, Kenney P, Manor M: Genetic parameter estimates for feed intake, body composition, and fillet quality traits in a rainbow trout population selected for improved growth. In: International Symposium on Genetics in Aquaculture. Auburn, AL; 2012: 259.
  17. Salem M, Al-Tobasei R, Ali A, Lourenco D, Gao G, Palti Y, Kenney B, Leeds TD: Genome-Wide Association Analysis With a 50K Transcribed Gene SNP-Chip Identifies QTL Affecting Muscle Yield in Rainbow Trout. Front Genet 2018, 9(387):387.
  18. Vallejo RL, Liu S, Gao G, Fragomeni BO, Hernandez AG, Leeds TD, Parsons JE, Martin KE, Evenhuis JP, Welch TJ et al: Similar Genetic Architecture with Shared and Unique Quantitative Trait Loci for Bacterial Cold Water Disease Resistance in Two Rainbow Trout Breeding Populations. Front Genet 2017, 8:156.
  19. Zheng X, Kuang Y, Lv W, Cao D, Sun Z, Sun X: Genome-Wide Association Study for Muscle Fat Content and Abdominal Fat Traits in Common Carp (Cyprinus carpio). PLoS One 2016, 11(12):e0169127.
  20. Ali A, Al-Tobasei R, Lourenco D, Leeds T, Kenney B, Salem M: Genome-Wide Association Study Identifies Genomic Loci Affecting Filet Firmness and Protein Content in Rainbow Trout. Frontiers in Genetics 2019, 10(386).
  21. Al-Tobasei R, Ali A, Leeds TD, Liu S, Palti Y, Kenney B, Salem M: Identification of SNPs associated with muscle yield and quality traits using allelic-imbalance analyses of pooled RNA-Seq samples in rainbow trout. BMC Genomics 2017, 18(1):582.
  22. Leeds TD, Vallejo RL, Weber GM, Pena DG, Silverstein JS: Response to five generations of selection for growth performance traits in rainbow trout (Oncorhynchus mykiss). Aquaculture 2016, 465:341-351.
  23. Hinshaw JM: Trout production: Feeds and feeding methods. . Southern Regional Aquaculture Center 1999, 223.
  24. Manor ML, Cleveland BM, Kenney PB, Yao J, Leeds T: Differences in growth, fillet quality, and fatty acid metabolism-related gene expression between juvenile male and female rainbow trout. Fish Physiology and Biochemistry 2015, 41(2):533-547.
  25. Zaitlen N, Kraft P, Patterson N, Pasaniuc B, Bhatia G, Pollack S, Price AL: Using extended genealogy to estimate components of heritability for 23 quantitative and dichotomous traits. PLoS Genet 2013, 9(5):e1003520.
  26. Liu S, Vallejo RL, Palti Y, Gao G, Marancik DP, Hernandez AG, Wiens GD: Identification of single nucleotide polymorphism markers associated with bacterial cold water disease resistance and spleen size in rainbow trout. Front Genet 2015, 6:298.
  27. Legarra A, Aguilar I, Misztal I: A relationship matrix including full pedigree and genomic information. J Dairy Sci 2009, 92(9):4656-4663.
  28. Misztal I, Tsuruta S, Lourenco D, Masuda Y, Aguilar I, Legarra A, Vitezica Z: Manual for BLUPF90 family of programs. Univ. Georg., Athens, USA; 2018.
  29. BLUPF90 and related programs (BGF90) [WWW Document], in Proceeding of 7th World Congress on Genetics Applied to Livestock Production (Montpellier) [ http://nce.ads.uga.edu/wiki/lib/exe/fetch.php?media=28-07.pdf]
  30. Turner SD: qqman: an R package for visualizing GWAS results using Q-Q and manhattan plots. bioRxiv 2014.
  31. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, Maller J, Sklar P, de Bakker PI, Daly MJ et al: PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 2007, 81(3):559-575.
  32. Song YE, Lee S, Park K, Elston RC, Yang HJ, Won S: ONETOOL for the analysis of family-based big data. Bioinformatics 2018, 34(16):2851-2853.
  33. Quinlan AR, Hall IM: BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 2010, 26(6):841-842.
  34. Lu D, Sargolzaei M, Kelly M, Vander Voort G, Wang Z, Mandell I, Moore S, Plastow G, Miller SP: Genome-wide association analyses for carcass quality in crossbred beef cattle. BMC Genet 2013, 14:80.
  35. Melody JL, Lonergan SM, Rowe LJ, Huiatt TW, Mayes MS, Huff-Lonergan E: Early postmortem biochemical factors influence tenderness and water-holding capacity of three porcine muscles. J Anim Sci 2004, 82(4):1195-1205.
  36. Zhang WG, Lonergan SM, Gardner MA, Huff-Lonergan E: Contribution of postmortem changes of integrin, desmin and μ-calpain to variation in water holding capacity of pork. Meat Sci 2006, 74(3):578-585.
  37. Garcia-Ruiz A, Cole JB, VanRaden PM, Wiggans GR, Ruiz-Lopez FJ, Van Tassell CP: Changes in genetic selection differentials and generation intervals in US Holstein dairy cattle as a result of genomic selection. Proc Natl Acad Sci U S A 2016, 113(28):E3995-4004.
  38. Thibeaux S, Siddiqi S, Zhelyabovska O, Moinuddin F, Masternak MM, Siddiqi SA: Cathepsin B regulates hepatic lipid metabolism by cleaving liver fatty acid-binding protein. J Biol Chem 2018, 293(6):1910-1923.
  39. Matsuo Y, Irie K, Kiyonari H, Okuyama H, Nakamura H, Son A, Lopez-Ramos DA, Tian H, Oka S, Okawa K et al: The protective role of the transmembrane thioredoxin-related protein TMX in inflammatory liver injury. Antioxid Redox Signal 2013, 18(11):1263-1272.
  40. Tandon P, Wafer R, Minchin JEN: Adipose morphology and metabolic disease. J Exp Biol 2018, 221(Pt Suppl 1).
  41. Joo JI, Yun JW: Gene expression profiling of adipose tissues in obesity susceptible and resistant rats under a high fat diet. Cell Physiol Biochem 2011, 27(3-4):327-340.
  42. Flor AC, Wolfgeher D, Wu D, Kron SJ: A signature of enhanced lipid metabolism, lipid peroxidation and aldehyde stress in therapy-induced senescence. Cell Death Discov 2017, 3:17075.
  43. Lemos MV, Chiaia HL, Berton MP, Feitosa FL, Aboujaoud C, Camargo GM, Pereira AS, Albuquerque LG, Ferrinho AM, Mueller LF et al: Genome-wide association between single nucleotide polymorphisms with beef fatty acid profile in Nellore cattle using the single step procedure. BMC Genomics 2016, 17:213.
  44. Söhle J, Machuy N, Smailbegovic E, Holtzmann U, Grönniger E, Wenck H, Stäb F, Winnefeld M: Identification of new genes involved in human adipogenesis and fat storage. PLoS One 2012, 7(2):e31193.
  45. Gaudet P, Livstone MS, Lewis SE, Thomas PD: Phylogenetic-based propagation of functional annotations within the Gene Ontology consortium. Brief Bioinform 2011, 12(5):449-462.
  46. Lin NY, Lin TY, Yang WH, Wang SC, Wang KT, Su YL, Jiang YW, Chang GD, Chang CJ: Differential expression and functional analysis of the tristetraprolin family during early differentiation of 3T3-L1 preadipocytes. Int J Biol Sci 2012, 8(5):761-777.
  47. Wahl S, Drong A, Lehne B, Loh M, Scott WR, Kunze S, Tsai PC, Ried JS, Zhang W, Yang Y et al: Epigenome-wide association study of body mass index, and the adverse outcomes of adiposity. Nature 2017, 541(7635):81-86.
  48. Wu YH, Ko TP, Guo RT, Hu SM, Chuang LM, Wang AH: Structural basis for catalytic and inhibitory mechanisms of human prostaglandin reductase PTGR2. Structure 2008, 16(11):1714-1723.
  49. Machnicka B, Czogalla A, Hryniewicz-Jankowska A, Bogusławska DM, Grochowalska R, Heger E, Sikorski AF: Spectrins: a structural platform for stabilization and activation of membrane channels, receptors and transporters. Biochim Biophys Acta 2014, 1838(2):620-634.
  50. Finocchiaro G, Ito M, Tanaka K: Purification and properties of short chain acyl-CoA, medium chain acyl-CoA, and isovaleryl-CoA dehydrogenases from human liver. J Biol Chem 1987, 262(17):7982-7989.
  51. Patwari P, Emilsson V, Schadt EE, Chutkow WA, Lee S, Marsili A, Zhang Y, Dobrin R, Cohen DE, Larsen PR et al: The arrestin domain-containing 3 protein regulates body mass and energy expenditure. Cell Metab 2011, 14(5):671-683.
  52. Lin JC: RBM4-MEF2C network constitutes a feed-forward circuit that facilitates the differentiation of brown adipocytes. RNA Biol 2015, 12(2):208-220.
  53. Mora S, Yang C, Ryder JW, Boeglin D, Pessin JE: The MEF2A and MEF2D isoforms are differentially regulated in muscle and adipose tissue during states of insulin deficiency. Endocrinology 2001, 142(5):1999-2004.
  54. Novokhatska O, Dergai M, Tsyba L, Skrypkina I, Filonenko V, Moreau J, Rynditch A: Adaptor proteins intersectin 1 and 2 bind similar proline-rich ligands but are differentially recognized by SH2 domain-containing proteins. PLoS One 2013, 8(7):e70546.
  55. Porat-Shliom N, Milberg O, Masedunskas A, Weigert R: Multiple roles for the actin cytoskeleton during regulated exocytosis. Cell Mol Life Sci 2013, 70(12):2099-2121.
  56. Komarova Y, De Groot CO, Grigoriev I, Gouveia SM, Munteanu EL, Schober JM, Honnappa S, Buey RM, Hoogenraad CC, Dogterom M et al: Mammalian end binding proteins control persistent microtubule growth. J Cell Biol 2009, 184(5):691-706.
  57. Wayne R, Tazawa M: The Actin Cytoskeleton and Polar Water Permeability in Characean Cells. In: Cell Dynamics: Molecular Aspects of Cell Motility Cytoskeleton in Cellular Structure and Activity. Edited by Tazawa M. Vienna: Springer Vienna; 1989: 116-130.
  58. Johnson JA, Hemnes AR, Perrien DS, Schuster M, Robinson LJ, Gladson S, Loibner H, Bai S, Blackwell TR, Tada Y et al: Cytoskeletal defects in Bmpr2-associated pulmonary arterial hypertension. Am J Physiol Lung Cell Mol Physiol 2012, 302(5):L474-484.
  59. Tan-Sindhunata MB, Mathijssen IB, Smit M, Baas F, de Vries JI, van der Voorn JP, Kluijt I, Hagen MA, Blom EW, Sistermans E et al: Identification of a Dutch founder mutation in MUSK causing fetal akinesia deformation sequence. Eur J Hum Genet 2015, 23(9):1151-1157.
  60. Cayrol C, Lacroix C, Mathe C, Ecochard V, Ceribelli M, Loreau E, Lazar V, Dessen P, Mantovani R, Aguilar L et al: The THAP-zinc finger protein THAP1 regulates endothelial cell proliferation through modulation of pRB/E2F cell-cycle target genes. Blood 2007, 109(2):584-594.
  61. Diamond G, Cedar H, Marcus M: A temperature-sensitive mutation in asparaginyl-tRNA synthetase causes cell-cycle arrest in early S phase. Exp Cell Res 1989, 184(1):53-60.
  62. Nakaseko Y, Yanagida M: Cell biology. Cytoskeleton in the cell cycle. Nature 2001, 412(6844):291-292.
  63. Bendris N, Lemmers B, Blanchard JM: Cell cycle, cytoskeleton dynamics and beyond: the many functions of cyclins and CDK inhibitors. Cell Cycle 2015, 14(12):1786-1798.
  64. Sontag E, Nunbhakdi-Craig V, Lee G, Bloom GS, Mumby MC: Regulation of the phosphorylation state and microtubule-binding activity of Tau by protein phosphatase 2A. Neuron 1996, 17(6):1201-1207.
  65. Qi H, Liu S, Guo C, Wang J, Greenaway FT, Sun MZ: Role of annexin A6 in cancer. Oncol Lett 2015, 10(4):1947-1952.
  66. Sugiyama M, Takenaga F, Kitani Y, Yamamoto G, Okamoto H, Masaoka T, Araki K, Nagoya H, Mori T: Homozygous and heterozygous GH transgenesis alters fatty acid composition and content in the liver of Amago salmon (Oncorhynchus masou ishikawae). Biol Open 2012, 1(10):1035-1042.
  67. Sferra A, Baillat G, Rizza T, Barresi S, Flex E, Tasca G, D'Amico A, Bellacchio E, Ciolfi A, Caputo V et al: TBCE Mutations Cause Early-Onset Progressive Encephalopathy with Distal Spinal Muscular Atrophy. Am J Hum Genet 2016, 99(4):974-983.
  68. Greenberg DL, Mize GJ, Takayama TK: Protease-activated receptor mediated RhoA signaling and cytoskeletal reorganization in LNCaP cells. Biochemistry 2003, 42(3):702-709.
  69. Huang L, Chardon JW, Carter MT, Friend KL, Dudding TE, Schwartzentruber J, Zou R, Schofield PW, Douglas S, Bulman DE et al: Missense mutations in ITPR1 cause autosomal dominant congenital nonprogressive spinocerebellar ataxia. Orphanet J Rare Dis 2012, 7:67.
  70. Hepler PK: The Cytoskeleton and Its Regulation by Calcium and Protons. Plant Physiol 2016, 170(1):3-22.
  71. Andersen KM, Madsen L, Prag S, Johnsen AH, Semple CA, Hendil KB, Hartmann-Petersen R: Thioredoxin Txnl1/TRP32 is a redox-active cofactor of the 26 S proteasome. J Biol Chem 2009, 284(22):15246-15254.
  72. David Y, Ziv T, Admon A, Navon A: The E2 ubiquitin-conjugating enzymes direct polyubiquitination to preferred lysines. J Biol Chem 2010, 285(12):8595-8604.
  73. Hatakeyama S, Yada M, Matsumoto M, Ishida N, Nakayama KI: U box proteins as a new family of ubiquitin-protein ligases. J Biol Chem 2001, 276(35):33111-33120.
  74. Ferreira V, van Dijk KW, Groen AK, Vos RM, van der Kaa J, Gijbels MJ, Havekes LM, Pannekoek H: Macrophage-specific inhibition of NF-kappaB activation reduces foam-cell formation. Atherosclerosis 2007, 192(2):283-290.
  75. Cahir-McFarland ED, Davidson DM, Schauer SL, Duong J, Kieff E: NF-kappa B inhibition causes spontaneous apoptosis in Epstein-Barr virus-transformed lymphoblastoid cells. Proc Natl Acad Sci U S A 2000, 97(11):6055-6060.
  76. Lin WJ, Yang CY, Li LL, Yi YH, Chen KW, Lin YC, Liu CC, Lin CH: Lysosomal targeting of phafin1 mediated by Rab7 induces autophagosome formation. Biochem Biophys Res Commun 2012, 417(1):35-42.
  77. Caillet M, Janvier K, Pelchen-Matthews A, Delcroix-Genête D, Camus G, Marsh M, Berlioz-Torrent C: Rab7A is required for efficient production of infectious HIV-1. PLoS Pathog 2011, 7(11):e1002347.
  78. Trombetta ES, Ebersold M, Garrett W, Pypaert M, Mellman I: Activation of lysosomal function during dendritic cell maturation. Science 2003, 299(5611):1400-1403.
  79. Jung J, Bohn G, Allroth A, Boztug K, Brandes G, Sandrock I, Schäffer AA, Rathinam C, Köllner I, Beger C et al: Identification of a homozygous deletion in the AP3B1 gene causing Hermansky-Pudlak syndrome, type 2. Blood 2006, 108(1):362-369.
  80. Zhu G, Ye R, Jung DY, Barron E, Friedline RH, Benoit VM, Hinton DR, Kim JK, Lee AS: GRP78 plays an essential role in adipogenesis and postnatal growth in mice. FASEB J 2013, 27(3):955-964.
  81. Zhu Z-L, Yang Q-M, Li C, Chen J, Xiang M, Chen M-M, Yan M, Zhu Z-G: Identification of biomarkers for childhood obesity based on expressional correlation and functional similarity. Molecular medicine reports 2018, 17(1):109-116.
  82. Hwangbo C, Wu J, Papangeli I, Adachi T, Sharma B, Park S, Zhao L, Ju H, Go GW, Cui G et al: Endothelial APLNR regulates tissue fatty acid uptake and is essential for apelin's glucose-lowering effects. Sci Transl Med 2017, 9(407).
  83. Yang Z, Bowles NE, Scherer SE, Taylor MD, Kearney DL, Ge S, Nadvoretskiy VV, DeFreitas G, Carabello B, Brandon LI et al: Desmosomal dysfunction due to mutations in desmoplakin causes arrhythmogenic right ventricular dysplasia/cardiomyopathy. Circ Res 2006, 99(6):646-655.
  84. Merline R, Schaefer RM, Schaefer L: The matricellular functions of small leucine-rich proteoglycans (SLRPs). J Cell Commun Signal 2009, 3(3-4):323-335.
  85. Writzl K, Maver A, Kovačič L, Martinez-Valero P, Contreras L, Satrustegui J, Castori M, Faivre L, Lapunzina P, van Kuilenburg ABP et al: De Novo Mutations in SLC25A24 Cause a Disorder Characterized by Early Aging, Bone Dysplasia, Characteristic Face, and Early Demise. Am J Hum Genet 2017, 101(5):844-855.
  86. Wei B, Yang Z, Cheng Y, Zhou J, Yang H, Zhang L, Yang X: Proteomic Analysis of the Hepatopancreas of Chinese Mitten Crabs (Eriocheir sinensis) Fed With a Linoleic Acid or  α-Linolenic Acid Diet. Front Physiol 2018, 9:1430.
  87. Ortega A, Roselló-Lletí E, Tarazón E, Molina-Navarro MM, Martínez-Dolz L, González-Juanatey JR, Lago F, Montoro-Mateos JD, Salvador A, Rivera M et al: Endoplasmic reticulum stress induces different molecular structural alterations in human dilated and ischemic cardiomyopathy. PLoS One 2014, 9(9):e107635.
  88. Pan Y, Cao F, Guo A, Chang W, Chen X, Ma W, Gao X, Guo S, Fu C, Zhu J: Endoplasmic reticulum ribosome-binding protein 1, RRBP1, promotes progression of colorectal cancer and predicts an unfavourable prognosis. British journal of cancer 2015, 113(5):763-772.
  89. Ponsuksili S, Jonas E, Murani E, Phatsara C, Srikanchai T, Walz C, Schwerin M, Schellander K, Wimmers K: Trait correlated expression combined with expression QTL analysis reveals biological pathways and candidate genes affecting water holding capacity of muscle. BMC Genomics 2008, 9:367.
  90. Miano JM, Long X, Fujiwara K: Serum response factor: master regulator of the actin cytoskeleton and contractile apparatus. Am J Physiol Cell Physiol 2007, 292(1):C70-81.
  91. Nicolas A, Kenna KP, Renton AE, Ticozzi N, Faghri F, Chia R, Dominov JA, Kenna BJ, Nalls MA, Keagle P et al: Genome-wide Analyses Identify KIF5A as a Novel ALS Gene. Neuron 2018, 97(6):1268-1283 e1266.
  92. Wu J, Zheng C, Wang X, Yun S, Zhao Y, Liu L, Lu Y, Ye Y, Zhu X, Zhang C et al: MicroRNA-30 family members regulate calcium/calcineurin signaling in podocytes. J Clin Invest 2015, 125(11):4091-4106.
  93. Xia HJ, Yang G: Inositol 1,4,5-trisphosphate 3-kinases: functions and regulations. Cell Res 2005, 15(2):83-91.
  94. Oh SW, Pope RK, Smith KP, Crowley JL, Nebl T, Lawrence JB, Luna EJ: Archvillin, a muscle-specific isoform of supervillin, is an early expressed component of the costameric membrane skeleton. J Cell Sci 2003, 116(Pt 11):2261-2275.
  95. Gross SR, Kinzy TG: Improper organization of the actin cytoskeleton affects protein synthesis at initiation. Mol Cell Biol 2007, 27(5):1974-1989.
  96. Salem M, Vallejo RL, Leeds TD, Palti Y, Liu S, Sabbagh A, Rexroad CE, 3rd, Yao J: RNA-Seq identifies SNP markers for growth traits in rainbow trout. PLoS One 2012, 7(5):e36264.
  97. Aguilar I, Legarra A, Cardoso F, Masuda Y, Lourenco D, Misztal I: Frequentist p-values for large-scale-single step genome-wide association, with an application to birth weight in American Angus cattle. Genet Sel Evol 2019, 51(1):28.
  98. Ali A, Al-Tobasei R, Lourenco D, Leeds T, Kenney B, Salem M: Genome-wide identification of loci associated with growth in rainbow trout. BMC Genomics 2020, 21(1):209.
  99. Lorenz AJ, Hamblin MT, Jannink JL: Performance of single nucleotide polymorphisms versus haplotypes for genome-wide association analysis in barley. PLoS One 2010, 5(11):e14079.

Tables

Due to technical limitations, Tables 1-5 are provided in the Supplementary Files sections.

CAPTIONS:

Table 1.  SNP markers in genomic sliding windows explaining at least 2% of the genetic variance for fat and moisture content and involved in lipid metabolism.  A color gradient on the left indicates differences in additive genetic variance explained by windows containing the representative SNP marker (green is the highest and red is the lowest).  SNPs are sorted according to their chromosome positions.

Table 2.  SNP markers in genomic sliding windows explaining at least 2% of the genetic variance for fat and moisture content and involvement in transmembrane transport and cytoskeleton regulation.  A color gradient on the left indicates differences in additive genetic variance explained by windows containing the representative SNP marker (green is the highest and red is the lowest).  SNPs are sorted according to their chromosome positions.

Table 3.  SNP markers in genomic sliding windows explaining at least 2% of the genetic variance for moisture content and involved in cell cycle and cytoskeleton regulation.  A color gradient on the left indicates differences in additive genetic variance explained by windows containing the representative SNP marker (green is the highest and red is the lowest).  SNPs are sorted according to their chromosome positions.

Table 4.  SNP markers in genomic sliding windows explaining at least 2% of the genetic variance for moisture content and involved in proteolytic activities.  A color gradient on the left indicates differences in additive genetic variance explained by windows containing the representative SNP marker (green is the highest and red is the lowest).  SNPs are sorted according to their chromosome positions.

Table 5. SNP markers significantly associated with variability in fat and moisture content using family-based association analysis.  A color gradient on the left shows the phenotypic variation explained by each SNP marker (green is the highest and red is the lowest).  SNPs associated with the phenotypes are sorted according to their chromosome positions.