Comparative evaluation of four reference genes for qRT-PCR assays in snow trout (Schizothorax richardsonii) under multiple experimental conditions Authors: Manchi Rajesh, Biju Sam Kamalam, Rini Joshi, Ciji Alexander, Md Shabaz Akhtar, Prakash Sharma Published: 2026 ABSTRACT AbstractThe analytical accuracy of gene expression analysis largely depends on the stability of the reference gene(s) used for normalization of target mRNA expression under different experimental conditions. In this study, we characterized three reference genes: beta-actin (β-act), elongation factor 1 alpha (ef1α), and glyceraldehyde 3-phosphate dehydrogenase (gapdh) in the Trans-Himalayan snow trout, Schizothorax richardsonii. Subsequently, we evaluated their stability, along with 18 S rRNA, under eight experimental conditions using the geNorm, NormFinder, BestKeeper, and delta CT algorithms. The results indicated that 18 S rRNA was the most stable reference gene across sixteen tissue types, whereas ef1α showed the highest stability during early developmental stages (0–21 days post-hatch) and between sexes (male and female muscle tissue). In other instances, β-act was the most stable reference gene in snow trout muscle across different age groups (0 + and 2 + years), nutritional conditions (starvation–refeeding and varying dietary protein and lipid levels), and rearing temperatures (6, 12, 18, and 24 °C). In contrast, gapdh exhibited poor stability and was regulated by low temperature, making it unsuitable as a reference gene under such conditions. For further validation, we performed relative quantification of primary myogenic regulators (myod and myf5) at different rearing temperatures using the most and least stable reference gene(s) as normalizers. The results reaffirmed the importance of selecting stable reference gene(s) for qRT-PCR assays, as the expression patterns of the myogenic regulators were significantly incorrect when normalized using the least stable reference genes. To our knowledge, this is the first comprehensive evaluation of internal reference genes for gene expression analysis in a coldwater cyprinid across different tissues, developmental stages, age, sex, nutritional conditions and temperature regimes.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-47431-8.Keywords: Gene expression analysis, Housekeeping gene, Nutrition, Temperature, Ontogeny, Age and sexSubject terms: Biological techniques, Genetics, Molecular biology, Physiology, Zoology Introduction Normalisation with a stable reference gene is essential in quantitative gene expression analysis of animal/plant tissues or microbial samples, using real-time qRT-PCR1, 2. The normalisation of gene expression is crucial to minimise potential errors introduced during sample processing and analysis, as it involves multiple steps like tissue sampling, homogenisation, RNA extraction, reverse transcription, dilution, and polymerase chain reaction. Often the differences in extraction efficiencies, reverse transcriptase or polymerase efficiency, and occasional carryover of inhibitors could introduce variations between samples. Also, the total RNA used for cDNA synthesis contains plentiful rRNA and the ratio of mRNA to rRNA can be perturbed under different experimental conditions3, affecting the actual mRNA quantity used for analysis. Normalization of gene expression data using internal reference controls or housekeeping genes (HKGs) helps to overcome the above-mentioned issues and ensures reliable results4. However, HKGs must be stably expressed across all samples and remain unaffected by the experimental conditions. For a given cell or tissue type, it is therefore to evaluate and identify stable reference genes prior to analysing target gene expression under specific experimental conditions for a given cell or tissue type, since universally stable HKGs do not exist1. With respect to fishes, Liao et al.5 summarised the studies on stability evaluation of reference genes under various experimental conditions such as developmental stages, environmental stress, pathogen infection, and exposure to toxic conditions in different tissue types. Among the many potential reference genes tested in fish, ef1α, 18 S rRNA, β-act, tubb2, tubb/a, ubq, rps4, and rpl17/28/32/8 were generally the most stable, but no single gene is universally stable across all tissues and experimental conditions⁵. Therefore, species-specific information on stably expressed reference genes under different nutritional, environmental and life-historical conditions is essential to support the growing interest in molecular biomarkers and gene expression studies in fish. The genus Schizothorax (family Cyprinidae) comprises nearly 45 species of snow trout distributed widely across the Trans-Himalayan region. These fishes are uniquely adapted to low- and high-altitude aquatic environments and exhibit distinct physiological responses to extreme environmental conditions6, 7, 8. Locally, they are highly valued for their taste and nutritional quality9. Among the snow trout species, Schizothorax richardsonii is considered particularly important for aquaculture diversification, especially in the context of unregulated exploitation of natural stocks, declining wild populations, and increasing consumer demand. Considerable scientific information is available on its breeding biology, reproductive physiology, and larval development10, 11. However, several biological and environmental constraints have been reported to limit its aquaculture potential, including inherently slow growth rates, susceptibility to parasitic infections, and reduced performance at elevated temperatures12–15. Addressing key knowledge gaps related to the snow trout’s biology, nutrition, environmental preferences, and health management is therefore essential for its sustainable aquaculture development, and modern biotechnological approaches offer promising opportunities in this regard. Furthermore, with the accelerating impacts of climate change in the Himalayan region, Schizothorax species have gained increasing recognition as sentinel organisms for studying climate resilience and adaptive physiological responses7. In recent years, genome editing has emerged as a powerful tool for improving economically important traits in fish species, particularly through the manipulation of muscle growth regulatory factors such as myogenic regulatory factors, myostatin, and follistatin15. Gene expression studies provide powerful insights into these biological processes; however, their reliability critically depends on the use of stable reference genes across muscle and other tissues. In the genus Schizothorax, only one study to date has evaluated a range of housekeeping genes namely 18 S rRNA, β-actin, ef1α, 40 S, b2m, tuba, ubce, and gapdh in non-muscle tissues following pathogen challenge16. However, several gene expression studies in Schizothorax richardsonii have used commonly employed housekeeping genes such as ef1α, β-actin, gapdh and 18 S rRNA for normalization without reporting their expression stability6, 11, 17, 18, 19. Hence, in this study, we cloned and sequenced three commonly used reference transcripts, namely β-act, ef1a, and gapdh in snow trout. Also, we assessed the stability of these candidate reference genes along with 18 S rRNA in snow trout in various tissue types, life-history (larval development, age, sex) differentiation, nutritional conditions (fed-starved status, dietary protein, and lipid levels), and rearing temperature regimes. Results Sequence information of snow trout β-act, ef1a and gapdh mRNA The characterized partial mRNA sequence of β-act, ef1a and gapdh in snow trout were 880 (Fig. S2, KU715835), 1345 (Fig. S3, KU715836) and 534 bp (Fig. S4, MH616628), respectively. The in silico observations are summarised in Table 1. Table 1Summary of sequence structure and Insilco analysis of sequenced candidate reference transcripts.GeneCDS5’ UTRCDS length3’ UTRPSN/O-GDomainsLocalizationInteracting partnersGBA/reference Sref1a Partial731272–YesYes–CytoplasmRP. EFγ, G-proteinKU715836/Current study Srβ-act Partial–1562589Yes–ActinCytoplasmMyH, ARP, Cofilin, ProfilinKU715835Current study Srgapdh Partial–534–Yes–GPDH-NPeriplasm, CytoplasmPGK, TES,Enolase, AldolaseMH616628Current studyCDS: Coding sequence, UTR: untranslated region; PS: phosphorylation sites; N/O-G: N/O- glycosylation sites; GBA: GenBank accession; EFγ: elongation factor 1 gamma; Gprotein: GTP binding protein; MyH: myosin heavy chain; ARP: actin related protein; PGK: Phosphoglycerate kinase 1; TES: Triosephosphate isomerase; E.S/R: extracellular space/region. Summary of sequence structure and Insilco analysis of sequenced candidate reference transcripts. KU715836/ Current study KU715835 Current study PGK, TES, Enolase, Aldolase MH616628 Current study CDS: Coding sequence, UTR: untranslated region; PS: phosphorylation sites; N/O-G: N/O- glycosylation sites; GBA: GenBank accession; EFγ: elongation factor 1 gamma; Gprotein: GTP binding protein; MyH: myosin heavy chain; ARP: actin related protein; PGK: Phosphoglycerate kinase 1; TES: Triosephosphate isomerase; E.S/R: extracellular space/region. Expression levels of candidate reference genes in snow trout The expression stability of four commonly used reference genes (β-act, ef1a, 18 S rRNA and gapdh) were studied across 16 different tissue-types, two life-history stages (early development and age class), sex and four experimental conditions (nutritional status, dietary protein and lipid intake levels, and different rearing temperature). Irrespective of experimental conditions, the expression level (Ct) of candidate reference gene β-act, ef1a, 18 S rRNA and gapdh in muscle tissue ranged from 21.1 to 22.9, 21.1–23.9, 8.1–10.9 and 19.6–21.8 respectively, in all the trials (Fig. 1C-H). Similar expression trend was also observed in whole larvae during the early developmental stages (Fig. 1B). However, between different tissues the Ct value of β-act, ef1a, 18 S rRNA and gapdh ranged from 16.0 to 22.9, 17.5–22.7, 8.3–11.4 and 20–34, respectively (Fig. 1A). The tissue-specific lowest and highest expression was noted in adipose and spleen in the case of β-act, adipose and spleen for ef1a, adipose and liver for 18 S rRNA, and brain and ovary for gapdh, respectively. Further, the pairwise variation in the expression of these four HKGs under different experiment conditions is presented in Figure S7. Violin plot depicting the expression levels of the four candidate reference mRNAs (18 S rRNA, β-act, ef1a and gapdh) under different experimental condition viz., across different tissues, early developmental time-course, age, sex, nutritional status, dietary protein levels (25, 35 and 45%), dietary lipid levels (3, 7 and 13%) and rearing water temperature (6, 12, 18 and 24 °C). The Ct variability is shown as range (whiskers) and median (horizontal line in the whiskers). Stability of the candidate reference genes across the snow trout tissues The median Ct values of the four candidate reference genes, namely β-act, ef1α, 18 S rRNA and gapdh, across sixteen different tissues were 18.61, 19.42, 9.2 and 26.35, respectively, whereas the coefficient of variation was 6.09, 4.55, 4.64 and 13.88%, respectively (Table ST1). All four algorithms provided a consensus result, ranking 18 S rRNA as the best reference gene, followed by ef1α and β-act (Table 2; Fig. 1A). According to NormFinder and geNorm, gapdh and β-act may not be suitable as normalizers for tissue-specific target gene expression due to their high standard deviation (SD > 1.0) and geNorm M values (M > 1.5), respectively. Furthermore, the geNorm M value indicated that 18 S rRNA and ef1α fall within the medium stability category (Fig S6A). As expected, the comprehensive ranking also reflected a similar stability order: 18 S rRNA > ef1α > β-act > gapdh (Table 3). Table 2Stability values of four candidate reference transcripts across different tissues, during early developmental time-course, age groups and sex.Algorithms/experimentsDifferent tissuesEarly development stagesβ-actef1a18 S rRNAgapdhβ-actef1a18 S rRNAgapdhDelta CT2.422.342.214.560.510.340.400.45Best Keeper1.140.890.433.710.360.190.220.40NormFinder1.331.160.574.490.420.260.260.34geNorm1.211.141.142.880.450.130.270.35Comprehensive ranking31.68143.7211.683.2Algorithms/experimentsAgeSexβ-actef1a18 S rRNAgapdhβ-actef1a18 S rRNAgapdhDelta CT0.670.710.710.930.670.630.661.05Best Keeper0.240.360.390.620.460.380.450.82NormFinder0.330.410.480.840.490.350.281.00geNorm0.480.480.580.750.210.210.450.75Comprehensive ranking1.321.862.0642.281.191.864 Stability values of four candidate reference transcripts across different tissues, during early developmental time-course, age groups and sex. Table 3Summary of the stability index of four candidate reference transcripts based on comprehensive ranking.ExperimentsStability ranking based on comprehensive ranking<< most stable ---- least stable >>Best combinations1.Different tissue 18SrRNA> ef1a > β-act - 2.Early developmental stage ef1a > 18 S rRNA > gapdh > β-act ef1a + 18SrRNA 3. Age/size β-act > ef1a > 18 S rRNA > gapdh β-act + ef1a 4.Sex ef1a > 18 S rRNA > β-act > gapdh ef1a + β-act 5.Nutritional status β-act > 18 S rRNA > ef1a > gapdh β-act + 18 S rRNA +ef1a 6.Dietary protein level β-act > ef1a > 18 S rRNA > gapdh β-act + ef1a + 18 S rRNA 7.Dietary lipid level β-act > 18 S rRNA > ef1a > gapdh β-act + 18 S rRNA+ ef1a 8.Rearing temperature β-act > ef1a > 18 S rRNA β-act + ef1a Summary of the stability index of four candidate reference transcripts based on comprehensive ranking. Stability of the candidate reference genes between early developmental stages, age and sexes of snow trout The median Ct values for β-act, ef1α, 18 S rRNA and gapdh during early larval developmental stages (larval stages) were 19.5, 19.95, 9.0 and 22.9, respectively, with coefficient of variations (CV) of 1.83, 0.93, 2.44 and 1.83%, respectively (Table ST1). All four algorithms and the comprehensive ranking consistently placed ef1α as the top reference gene, followed by 18 S rRNA, gapdh, and β-act. However, BestKeeper ranked β-act as the third-best reference gene, followed by gapdh (Table 2). This could be due to a slightly lower standard deviation, even though the gapdh had a lower coefficient of variation (Table 2). Overall, all the four candidate genes exhibited characteristics of being good reference genes with a standard deviation of less than 0.5 and high stability (geNorm M value less than 0.5). With respect to the two age groups (0 + and 2+), the median Ct values for β-act, ef1α, 18 S rRNA and gapdh in muscle tissue were 22.10, 22, 8.4 and 21.83, with coefficients of variation (CV) of 1.78, 1.63, 2.90 and 2.85%, respectively (Table ST1). All the algorithms consistently ranked β-act as the most stable gene, followed by ef1α and 18 S rRNA, while gapdh was found to be the least stable reference gene (Table 2). Based on the geNorm analysis, β-act and ef1α can be classified as highly stable reference genes, while 18 S rRNA and gapdh can be considered as moderately stable internal control genes for normalizing target gene expression between different age groups of snow trout (Fig. S6C). The median Ct values for the candidate reference transcripts, namely β-act, ef1α, 18S rRNA and gapdh between male and female snow trout were 22.63, 23.1, 9.58 and 20.59, respectively. The coefficients of variation for these transcripts were 2.06, 1.64, 4.74 and 3.94%, respectively (Table ST1). Three out of the four algorithms consistently ranked ef1α as a stable reference gene, followed by 18S rRNA and β-act, while gapdh was considered the least stable among the four (Table 2). However, it is worth noting that NormFinder ranked 18S rRNA as the best stable gene over ef1α (Table 2). This discrepancy may be attributed to the slightly lower variations in both inter and intra-group variations in 18S rRNA expression, despite a higher CV of 4.74% (Table 2), which could be due to its abundant expression pattern. On the other hand, geNorm categorized ef1α and β-act as highly stable reference genes and 18S rRNA and gapdh as moderately stable reference genes. According to geNorm’s ‘V’ value, ef1α and β-act represent the best combination of stable reference genes among the four (Fig. S6D). Stability of the candidate reference gene expression in snow trout under different feeding regimes The stability of candidate reference genes was analysed in muscle tissue of snow trout subjected to different feeding regimes: 21 days of regular feeding, 21 days of feed deprivation, and short-term (2 days) and long-term (21 days) refeeding after 3-week starvation. The median Ct values for the genes β-act, ef1α, 18 S rRNA and gapdh across these four treatment groups were: 21.12, 21.38, 8.26 and 20.73, with CV of 3.47, 3.60, 2.93 and 2.52%, respectively (Table ST2). The delta Ct and geNorm analyses ranked β-act and ef1α as the stable reference genes, followed by 18 S rRNA and gapdh (Table 4). It is worth noting that these algorithms assess stability based on pairwise variation, and there was a significant correlation between the expression patterns of β-act and ef1α (Pearson correlation coefficient, r = 0.938, p = 0.001). On the other hand, BestKeeper ranked the reference genes in the order of 18 S rRNA > gapdh > β-act > ef1α due to lower standard deviation and CV in 18 S rRNA and gapdh (Table ST2). NormFinder ranked them in the order of β-act > 18 S rRNA > ef1α > gapdh, based on inter and intra-group variations in the Ct values of these candidate reference genes. Overall, it is important to note that all four genes had SD values less than 1 and geNorm M = 0.5–1.5, placing them in the medium stability category. Additionally, geNorm analysis suggested that the combination of the three best-performing genes, namely β-act, ef1α, and 18 S rRNA, could be a suitable choice for normalizing expression analysis of the target gene under starvation and refeeding conditions (Fig. S6E). Table 4Stability values of four candidate reference transcripts under different experimental conditions viz. nutritional status, dietary protein levels, dietary lipid levels and rearing temperatures.Algorithms/experimentsNutritional statusDietary protein levelsβ-actef1a18 S rRNAgapdhβ-actef1a18 S rRNAgapdhDelta CT0.880.920.961.070.470.50.520.69Best Keeper0.730.760.240.520.370.40.50.56NormFinder0.610.70.640.890.250.350.280.63geNorm0.550.590.630.760.220.220.390.54Comprehensive ranking1.322.212.063.3611.862.714Algorithms/experimentsDietary lipid levelsDifferent rearing temperaturesβ-actef1a18 S rRNAgapdhβ-actef1a18 S rRNAgapdhDelta CT0.40.440.410.671.111.151.172.3Best Keeper0.280.310.160.450.440.480.681.75NormFinder0.200.280.120.640.560.680.342.26geNorm0.230.230.290.480.260.260.571.43Comprehensive ranking1.412.281.5741.191.862.284 Stability values of four candidate reference transcripts under different experimental conditions viz. nutritional status, dietary protein levels, dietary lipid levels and rearing temperatures. Stability of candidate reference genes expression in snow trout fed different levels of dietary protein The stability of the four candidate reference genes was examined in the muscle tissue of snow trout fed diets with 25, 35 and 45% protein levels. The median Ct values for the genes β-act, ef1α, 18 S rRNA and gapdh across these three treatment groups were 22.8, 23.4, 9.4 and 20.6, with coefficients of variation of 1.61, 1.72, 5.22 and 2.73%, respectively (Table ST2). Three out of the four algorithms and comprehensive ranking indicated β-act and ef1α as the most stably expressed genes, followed by 18 S rRNA and gapdh (Table 4). However, NormFinder ranked the stability of gene expression in the following order: β-act > 18 S rRNA > ef1α > gapdh. In geNorm β-act, ef1α, and 18 S rRNA were classified as highly stable reference genes (M < 0.5), while gapdh was of medium stability (M = 0.54). Additionally, geNorm V values showed that, a combination of β-act and ef1α was found to be sufficient for reliable mRNA expression analysis of the target genes in snow trout under different dietary protein intake levels (Fig. S6F). Stability of candidate reference genes expression in snow trout fed different levels of dietary lipid In another experiment, the stability of reference genes was assessed in the muscle tissue of snow trout fed diets with three different levels of lipids. The median Ct values for the genes β-act, ef1α, 18 S rRNA and gapdh across these three treatment groups were 22.1, 22.83, 8.1 and 19.65 with coefficients of variation of 1.27, 1.37, 1.97 and 2.28%, respectively (Table ST2). BestKeeper and NormFinder algorithms ranked the stability of expression of candidate reference genes as 18 S rRNA > β-act > ef1α > gapdh (Table 4). On the other hand, Delta Ct and geNorm algorithms ranked β-act, 18 S rRNA and ef1α as the most stable genes (M < 0.5), while gapdh as a gene of medium stability (M = 0.5). According to the geNorm V value, a combination of only two genes (β-act and ef1α, V < 0.15) were sufficient for acceptable normalization of gene expression (Fig. S6G). Overall, the comprehensive ranking of the four candidate reference genes under different dietary lipid levels was β-act > 18 S rRNA > ef1α > gapdh (Table 3). Stability of candidate reference genes expression in snow trout reared at different thermal regimes The median Ct values for the four candidate reference genes, namely β-act, ef1α, 18 S rRNA and gapdh, in the muscle tissue of snow trout at different rearing temperatures were 22.8, 23.9, 10.65 and 21.05 with a CV of 1.93, 2.01, 6.33 and 8.01%, respectively (Table ST2). Three out of the four algorithms identified β-act as the most stable reference gene, followed by ef1α and 18 S rRNA (Table 4). But, NormFinder ranked 18 S rRNA as the most stable reference gene, as it had an SD of 0.68 despite the high CV. On the other hand, gapdh was found to be least stable gene, as it showed a high standard deviation (SD = 1.75) and a geNorm M value greater than 1.5 (Fig. S6H). Overall, the comprehensive ranking suggests a stability order of β-act > ef1α > 18 S rRNA (Table 3). Validation of candidate reference genes for normalization The expression patterns of the target genes showed that both myod and myf5 expressions were sensitive to the choice of normalizer (Fig. 2). Normalizing the target genes with the most stable reference genes (β-act and ef1α) resulted in a growth-correlated expression pattern of the primary myogenic regulator (Fig. 2A and B). On the other hand, normalizing with the medium to least stable normalizers, 18 S rRNA and gapdh, resulted in either no change or a completely different expression pattern. Likewise, using a combination of β-act and ef1α yielded similar results as that of the best individual reference genes (Fig. 2C), but introducing a third reference gene (18 S rRNA) affected the statistical significance of the target gene expression (Fig. 2D). The combination of all four reference genes did not result in satisfactory expression results for myf5 and myod, due to the poor stability of gapdh across different rearing temperatures (Fig. 2E). Validation of reference genes: Relative gene expression of primary myogenic regulators (A) Srmyf5 and (B) Srmyod in snow trout under different rearing temperature, normalized using two most stable (β-act and ef1α) and least stable (18 S rRNA and gapdh) transcripts. (C) Expression of Srmyf5 and Srmyod normalised using geomean of two reference transcripts (β-act and ef1α). (D) Expression of Srmyf5 and Srmyod normalised using geomean of three reference transcripts (β-act, ef1α and 18 S rRNA). (E) Expression of Srmyf5 and Srmyod normalised using geomean of all four reference transcripts (β-act, ef1α, 18 S rRNA and gapdh). All data are presented as means ± s.d. (n = 6) and were analysed using one way ANOVA followed by Tukey’s multiple comparison test. Bars labelled with different superscript letters (a, b) indicate significant differences between treatments (p < 0.05). Discussion The objective of this study was to characterize and validate the major candidate internal reference genes, namely β-actin (β-act), elongation factor 1 alpha (ef1α), glyceraldehyde 3-phosphate dehydrogenase (gapdh), and 18 S ribosomal RNA (18 S rRNA) for real-time qRT-PCR assays in snow trout (Schizothorax richardsonii). The stability of these reference genes were assessed across different tissue types, larval developmental stages, age, sex, nutritional conditions (including starvation, refeeding, and varying dietary protein and lipid levels), and thermal regimes. These four candidate genes were selected because they are the most used reference genes representing different functional classes. Ef1α plays a key role in protein synthesis by facilitating the decoding of mRNA codons into amino acids, a fundamental cellular process that occurs continuously, and is therefore expected to be constitutively expressed20. Similarly, 18 S rRNA is a structural component of ribosomes and is directly involved in the process of protein synthesis. It is typically expressed at high levels, often with Ct values below 12, making it suitable for normalizing highly expressed transcripts21. Beta-actin is an essential cytoskeletal protein that plays a critical role in maintaining cell structure, integrity, and motility, and it is frequently used as a reference gene because of its relatively stable expression pattern across tissues and conditions22, 23. The gapdh is an important enzyme in the glycolytic pathway, it also plays non-metabolic roles such as apoptosis induction, receptor-associated kinase activity, DNA repair, and axonal transport24, 25. It is widely used as an internal control for gene expression analysis from fish to humans26, 27. In the current study, we used four algorithms to evaluate the expression stability and all four algorithms generally ranked the most and least stable genes consistently, except BestKeeper and NormFinder, which ranked 18 S rRNA as the best gene in 3 out of 8 experiments (Tables 2 and 4). This discrepancy may be due to the smaller numerical standard deviation of 18 S rRNA, linked to its abundant expression pattern (median Ct = 8.4) compared to protein-coding genes (median Ct ~ 22) (Fig. 1 and Table ST1 and ST2). When candidate reference genes exhibit higher standard deviation in expression levels (Ct values), it may be useful to consider the coefficient of variation (CV) along with SD to rank gene stability. For example, in different age groups, 18 S rRNA had a SD of 0.24 and a CV of 2.9%, while β-act had an SD of 0.39 and a CV of 1.39% (Table ST1). Similar discrepancies among algorithms have been observed in other studies5, 21, 28, 29, highlighting the importance of using multiple algorithms to select stable reference genes from a panel of candidates. When the stability of candidate transcripts across sixteen tissues of snow trout was analysed, 18 S rRNA was found to be most stable among the four (Table 2). Coherent with the current finding, earlier studies in spotted seabass, Nile tilapia, zebrafish, channel catfish and Misgurnus anguillicaudatus had reported that, 18 S rRNA is a suitable internal control for gene expression analysis across different tissue samples21, 30, 31, 32, 33, 34. However, it should be noted that 18 S rRNA ranked as a medium stability category gene by geNorm which is typically observed in a heterogeneous set of samples, such as different tissue types1. Gene expression analysis during larval developmental stages is crucial for a better understanding of fish developmental biology, particularly concerning organ system development such as the digestive and immune systems. Identifying stably expressing reference gene is a prerequisite for such gene expression assays. Therefore, we investigated the stability of four commonly used HKGs in snow trout over the time course of larval development (0 to 21 days post-hatching). The results indicated that, ef1α consistently ranked as the most stable gene across all four algorithms, during the early developmental stage (Table 2). Interestingly, all four genes exhibited stable expression patterns, with indices below the threshold values in all algorithms (Table 2). Furthermore, the correlation of ef1α, 18 S rRNA, and gapdh resulted in r ≥ 0.8 (Pearson correlation coefficient, p < 0.05; BestKeeper Index), indicating that any one of these three genes or their combinations could serve as effective normalization factors. According to the geNorm V value, the optimal number of reference genes is two (ef1α and 18 S rRNA, V = 0.116), which is sufficient to obtain a satisfactory expression data for target transcripts during early development. No further addition or combination of genes significantly reduced variation (Fig S6B and S7B). Several studies have similarly reported that ef1α is stably expressed across developmental stages in various fish species, including Atlantic salmon, weather loach, Atlantic halibut, European seabass, Senegalese sole, Asian seabass, and striped murrel29, 34, 35, 36, 37, 38, 39. In other fish species like Catla catla and Clarias magur (catfish), gapdh and β-act were found to be the most stably expressed reference genes during early developmental phase40, 41. This difference in reference gene expression patterns could be attributed to species-specific variations in developmental rates, such as precocial and altricial development patterns in different fish, and the specific developmental stages at which samples were collected. Further, the stability of reference genes was assessed in two age groups (0 + and 2 + years) of snow trout. The stability ranking order was β-act > ef1α > 18 S rRNA > gapdh (Tables 2 and 3). β-act and ef1α were identified as the most stably expressed reference genes (geNorm ≤ 0.15) compared to the other two. This finding aligns with a previous study in zebrafish, where β-act showed the lowest expression variability from larval to adult stages among the eight HKGs tested42. Similarly, a study in juvenile and adult salmon established that ef1α was the second least varying gene among the four HKGs tested43. When the stability of reference genes was analysed between sexes (male and female), the comprehensive algorithm ranked ef1α as the most stable reference gene (ef1α > β-act > 18 S rRNA > gapdh) (Tables 2 and 3). Consistent with our findings, ef1α was also identified as the most stable reference gene in zebrafish and Misgurnus anguillicaudatus, while gapdh expression was highly variable31, 34. In addition, ef1α and β-act could be used in combination for normalizing gene expression in different sexes, as indicated by the geNorm algorithm. Notably, there was a strong positive correlation in the expression patterns of ef1α and β-act (Pearson correlation coefficient, r = 0.936, p = 0.01). Nutrition and feeding trials are often conducted to understand the physiological responses of fish to dietary changes and develop optimal feeds for different species. In these studies, the expression patterns of genes involved in growth regulation, digestive enzymes, metabolism, and the immune system are quantified to elucidate molecular mechanisms. Accurate normalization of the target gene expression results requires knowledge of stably expressed reference genes under different nutritional regimes. In the current study, we evaluated the stability of reference genes in snow trout muscle tissue across three feeding trials - nutritional status, different dietary protein, and lipid levels. The results indicated that β-act was the most stably expressed reference gene in the three nutritional studies (Table 4). Consistent with this finding, an earlier study in tiger puffer also reported β-act as the most stable reference transcript under varying nutritional conditions, such as starvation and different dietary lipid levels5, 32. Similarly, β-act was found to be the most suitable stable reference gene in hybrid yellow catfish fed different levels of soybeans44. However, it should be noted that the stable reference gene may vary with tissue-type, as ef1a was reported to be stably expressed in the liver tissue of walking catfish and largemouth bass45, 46. The similarities and differences in the stability of reference genes across species, experimental conditions, and tissue types underscores the importance of selecting and validating reference genes for specific experimental conditions. Temperature is a critical factor that governs the metabolism and growth rate of ectotherms like fish47. In their natural habitat, snow trout experience pronounced seasonal temperature fluctuations ranging from 0 to 27 °C7. Due to this wide thermal range, snow trout represent a valuable model for investigating the molecular mechanisms underlying temperature tolerance and adaptation to climate change in mid and high-altitude environments6, 7. In this context, we investigated the stability of four reference genes in snow trout muscle across four rearing temperatures (6, 12, 18 and 24 °C). The results indicated that β-act was the least variable reference gene, with the overall stability ranking of β-act > ef1α > 18 S rRNA (Table 3). This finding is consistent with previous studies reporting β-act as the most stable reference gene in the muscle tissue of striped murrel exposed to elevated temperatures48. Moreover, β-act has also been successfully used as a reference gene for temperature-related gene expression studies in the closely related species Schizothorax prenanti49. On the other hand, gapdh appeared to be unsuitable for normalizing muscle gene expression in snow trout under varying temperature conditions. In the present study, gapdh showed significant downregulation at lower temperatures, which may reflect a reduction in metabolic rate, similar to a previous report in African lungfish50. Besides, the stability of reference genes can be tissue-specific, as 18 S rRNA, ef1α and rps29/rpl19 were found to be the most stable reference genes in non-muscle tissues of Atlantic cod, Puntius sophore, and Hucho taimen, respectively, under short-term and long-term temperature stress51, 52, 53, 54. Finally, to emphasize the importance of selecting a stable reference gene for analysing target gene expression, we normalized the gene expression data using the best and least stable reference genes, as well as their combinations. In this validation experiment, we examined the expression of two primary myogenic regulators, myod and myf5 in the snow trout muscle under different rearing temperatures. The myogenic regulatory factors play crucial roles in muscle cell determination, differentiation, and growth55. This experiment was chosen due to the significant phenotypic responses (growth) observed across four rearing temperatures17. Normalization using the stable reference genes β-act and ef1α, either individually or in combination, produced consistent expression patterns (Fig. 2). The upregulation of myf5 corresponded with the maximum growth increment observed at 18 °C. Similar correlations between primary myogenic regulators and growth responses have been observed in rainbow trout and grass carp56, 57. Conversely, normalization with the least stable reference gene (18 S rRNA and gapdh) resulted in contradictory expression patterns of myogenic regulators which did not reflect growth phenotype (Fig. 2). This validation experiment clearly demonstrates and reaffirms the necessity of screening for stable reference genes under the specific testing conditions for a given species or tissue type, to ensure reproducible and accurate gene expression assays. Based on the validation results from this study, β-act and ef1α can be recommended as the most stable reference genes for temperature-related gene expression studies in the muscle tissue of snow trout and potentially for other coldwater cyprinids. It is important to note that normalization using more than one reference gene is also advisable, as it helps reduce variability and enhances the accuracy of qRT-PCR results. Conclusion In conclusion, through this comprehensive study, we characterized and validated the stability of major candidate reference genes in the Himalayan snow trout, across different experimental conditions. Our findings showed that 18 S rRNA was the most stable reference gene across different tissue types; ef1α was the most stable HKG during early development and between male and female snow trout; and β-act was the most stable reference gene across age groups, nutritional regimes, and rearing temperatures. This also reaffirms that there is no universally stable reference gene across experimental conditions and tissue types, even within the same species. In addition, our validation study highlighted the importance and necessity of selecting and using a stable reference gene for gene expression assays. Overall, the findings of this study provide a critical foundation for future transcriptomic and qPCR-based studies on diverse physiological aspects of snow trout and other Himalayan coldwater cyprinids to support the sustainable aquaculture and conservation. Materials and methods Ethical statement All the experimental procedures involving snow trout (Schizothorax richardsonii) were reviewed and approved by the Institutional Animal Care and Use Committee of ICAR–Central Institute of Coldwater Fisheries Research, Bhimtal, India (Approval No. DCFR/IACUC/12/06/2020/4). Fish were maintained and handled in accordance with the guidelines of the Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA), Ministry of Environment, Forest and Climate Change, Government of India. Every effort was made to minimize pain, stress, and suffering of the experimental fish. Fish were deeply anesthetized with clove oil (100 µL L⁻¹) prior to all invasive procedures and were humanely euthanized by needle-assisted cervical dislocation as a secondary physical method to ensure rapid death. Amplification and cloning of β-act, ef1α and gapdh in snow trout In the current study, three housekeeping genes, β-act, ef1α and gapdh were cloned and sequenced. Additionally, 18 S rRNA was included as a candidate reference gene for stability analysis. The sequence information for 18 S rRNA of this species was obtained from GenBank (KX235299). To obtain the nucleotide sequence information of reference mRNA (β-act, ef1a and gapdh) in snow trout, initially partial gene characterisation of β-act and gapdh was carried out using the total RNA isolated from muscle tissue of an adult fish (using Trizol). The quantity and quality of the extracted RNA were evaluated using spectrophotometric method (Biotech Eon3 VT, USA) and agarose gel electrophoresis, respectively. The cDNA synthesis was carried out using Revertid H Minus Reverse Transcriptase and oligo dT primer, as per the manufacturer’s protocol (Thermo Fisher Scientific, MA, USA). Degenerate or gene specific forward and reverse primers (Table 5) were designed targeting the conserved regions of cyprinid β-act and gapdh, using Clustal Omega and IDT OligoAnalyzer online tool. The partial amplification of β-act and gapdh was carried out using High-Fidelity Phusion hot start polymerase enzyme as per the previously described protocol17. Amplified products were cloned in TOP10 competent Escherichia coli using pCR-Blunt vector (Invitrogen, Carlsbad, CA, USA). Plasmids were extracted from positive clones and sequenced in triplicates using universal primers (T7 and T3). Table 5Primers used for partial and RACE PCR amplification of candidate reference genes in snow trout.PrimersSequence 5’−3’For PCR β-act F-GGTATGGGACAGAAGGACAGR-TGGAAGGTGGACAGGGAG gapdh F-CATGGTTTACATGTTCAAGTATGACTCR-GACTTTGCCCACAGCCTTGGCFor RACEβ-act 3’RACE-1GGTCATCACCATTGGCAATGAGAGGTTCAGef1a 5’RACE-1CTCACCAACACCACCAGCAACAATCAGCACef1a 5’RACE-2GCTGACTTCCTTGGTGATTTCCTCAAAACGAnchor oligodT3’ IHGAGTGTCGCACGGTCCATCGCAGCTTTTTTTTTTTTTTTTTTTTTTTTTTVN3’RACE long-IHCGCTGACGCTGAGCCTACCTGACGAGTGTCGCACGGTCCATCGCAG3’ RACE short-IHCGCTGACGCTGAGCCTACCTGACFISH TSO C3 spacer-IHGAGTGTCGCACGGTCCATCGCAGCAGTCACAGGGGG/3SpC3/5’RACE USE primerGAGTGTCGCACGGTCCATCGCAGC Primers used for partial and RACE PCR amplification of candidate reference genes in snow trout. F-GGTATGGGACAGAAGGACAG R-TGGAAGGTGGACAGGGAG F-CATGGTTTACATGTTCAAGTATGACTC R-GACTTTGCCCACAGCCTTGGC The 3′ and 5′ ends of β-act and ef1α (ef1α was partially characterized in our previous study58 were amplified using an in-house rapid amplification of cDNA ends (RACE) method. The in-house 5′ RACE protocol was modified and adapted from Pinto and Lindblad59 to suit eukaryotic RNA, specifically RNA extracted from fish samples. This protocol enables the amplification of specific and complete 5′ untranslated regions (UTRs) while avoiding the step-out PCR and subsequent nested PCR steps commonly required in many commercial kits. This is achieved by adding Mn²⁺ ions to the reaction. Mn²⁺ enhances the terminal cytosine addition activity of Murine Leukemia Virus reverse transcriptase upon recognition of the 5′ cap structure (7-methylguanosine). Specificity is further ensured by modification of the template-switching oligo (TSO) primer through 3′ blocking59. In addition, this method is cost-effective, as sufficient probes can be synthesized for multiple reactions; whereas, commercial kits typically support only 5–10 reactions. The detailed procedure followed for 5′ RACE is provided in the Supplementary Materials (Fig. S1). Experimental fish and sampling procedure The experimental snow trout were collected from Kalsa stream, Chaffi, Uttarakhand, India. For feeding trials, fishes were maintained in ICAR-CICFR wet laboratory facility in 1 m3 rectangular flow through tanks, and for age-sex-tissue studies fishes were immediately euthanized and sampled. During the acclimation period, they were fed a compound diet containing 35% protein and 8% lipid. For all the feeding trials, the well acclimatised (after ~ 3 months in captivity) fish were randomly allotted to different tanks (110 L rectangular tanks) in either static (nutritional status) flow-through (dietary protein) or recirculating aquaculture system (dietary lipid and temperature). At the beginning and end of the trial, fish weights were recorded to the closest gram weight. Fish were anesthetised with clove oil (100 µL/L, CDH, India) and humanely euthanized by cervical dislocation at the conclusion of the corresponding studies (Table 6). Tissue samples were aseptically dissected out, snap frozen in liquid nitrogen and preserved at −80 °C for pending analysis. Table 6Summary of the different experiments performed in S. richardsonii to evaluate the stability of potential reference genes.Sr. no.ExperimentsExperimental conditionsDuration of experiment (days)System1Early development stage, n = 65,7,9,12,15,18 and 21 days post hatching–Flow through hatchery2Age/size groupsn = 60+(3.2 ± 0.6 g)2+(82.3 ± 13.4 g)–Wild collected3Sexesn = 6 (33.3 ± 6.9 g)MaleFemale–Wild collected4Different tissue(n = 4;Two ♂and ♀ each).16 different tissues namely anterior dorsal muscle, posterior caudal muscle, liver, kidney, eye, adipose tissue, skin, spleen, heart, gill, brain, foregut, midgut, hindgut, testis, and ovary.5Nutritional status,n = 6 (~ 5 g)Fed(21 days)Starved(21 days)ST-RF(2 days)LT-RF(21 days)-Static experimental setups, (9 tanks)6Dietary protein levelsn = 6 (~ 25 g)25%35%45%100RAS-based experimental unit, 9 tanks7Dietary lipid levelsn = 6 (~ 20 g)3%7%13%68RAS-based experimental unit, 9 tanks8Rearing temperaturen = 6 (~ 10 g)6 °C12 °C18 °C24 °C75RAS-based experimental unit, (4 tanks)‘n’ indicates the number of biological replicates used per treatment for reference gene expression analysis. ST-RF: short-term refeeding (2 days of refeeding following 21 days of starvation); LT-RF: long-term refeeding (21 days of refeeding following 21 days of starvation). For tissue expression analysis, n = 4 for all tissues except testis and ovary, where n = 2. Summary of the different experiments performed in S. richardsonii to evaluate the stability of potential reference genes. Age/size groups n = 6 0+ (3.2 ± 0.6 g) 2+ (82.3 ± 13.4 g) Sexes n = 6 (33.3 ± 6.9 g) Different tissue (n = 4; Two ♂and ♀ each). Nutritional status, n = 6 (~ 5 g) Fed (21 days) Starved (21 days) ST-RF (2 days) LT-RF (21 days) Dietary protein levels n = 6 (~ 25 g) Dietary lipid levels n = 6 (~ 20 g) Rearing temperature n = 6 (~ 10 g) ‘n’ indicates the number of biological replicates used per treatment for reference gene expression analysis. ST-RF: short-term refeeding (2 days of refeeding following 21 days of starvation); LT-RF: long-term refeeding (21 days of refeeding following 21 days of starvation). For tissue expression analysis, n = 4 for all tissues except testis and ovary, where n = 2. Experimentations Detailed information about each feeding trials and sampling methods were provided in Rajesh et al.17 and summarised here in Table 6. In all the wet lab experiments, the water temperature was maintained between 18–20℃ (except for experiment 8) and other water quality metrics like pH, dissolved oxygen (DO), carbon dioxide (CO2), alkalinity, total ammonia nitrogen (TAN) and nitrite were frequently examined and confirmed to be within safe levels. In the 8th experiment, different rearing temperatures 6, 12, 18 and 24℃ were maintained using either an aquarium chiller (Hailea HC-1000 A, China; 6 °C) or 300 W metallic aquarium heaters (12, 18 and 24 °C) connected to a thermostat. Primer design for reference and target genes Once the nucleotide sequences for the candidate reference (β-act, ef1a, gapdh and 18 S rRNA) and target genes (myf5 and myod) were identified, qPCR primers were designed following the method described by Thornton and Basu60 using Primer 3.0 software (http://primer3.ut.ee/). The primer parameters and secondary structures were analysed using the IDT OligoAnalyzer and Beacon Designer online tool. Several sets of primers were designed, and only those that met acceptable in silico primer properties were selected for synthesis. After synthesizing the primers, they were validated by qPCR screening for amplification efficiency and specificity. Primer dimer formation was checked using agarose gel electrophoresis and melt curve analysis (Fig. S5). Only primers with an amplification efficiency of 90–110%, a single peak melt curve, and no primer dimer formation were selected for further gene expression analysis (Table 7). Table 7Real time qPCR primers used for the validation of reference genes in snow trout.PrimersSequence 5’−3’Ta(oC)Amplification efficiency%Amplicon size (bp)GenBank ID Ef1a-rt F- ACAAACATGGGCTGGTTCAAGGR- AAATGGCATCCAGGGCATCAAG60107.291KU715836 βact-rt F- AGAAGAGCTATGAGCTGCCTGAR- ATACCCAAGAAGGATGGCTGGA60100.396KU715835 Gapdh-rt F- AGTTGTCAGCAATGCCTCCT6097140MH616628R- GGTCTTCTGTGTGGCTGTGA 18 S-rt F- AAACGGCTACCACATCCAAGGR- ATGGGTTTAGGATACGCTCATTCC60100.5135KX235299 Myf5-rt F9- CGGAGAGCCGCCACTATGAR8- CCACCTTGGGAAGACGCTGA6096.3111OP272172 Myod-rt F1- CCATGTGGCTCCCAATGAGTATCR1- ACTCCACTGAGGCTTCTTCTC60103.3113MH616627 Real time qPCR primers used for the validation of reference genes in snow trout. F- ACAAACATGGGCTGGTTCAAGG R- AAATGGCATCCAGGGCATCAAG F- AGAAGAGCTATGAGCTGCCTGA R- ATACCCAAGAAGGATGGCTGGA F- AAACGGCTACCACATCCAAGG R- ATGGGTTTAGGATACGCTCATTCC F9- CGGAGAGCCGCCACTATGA R8- CCACCTTGGGAAGACGCTGA F1- CCATGTGGCTCCCAATGAGTATC R1- ACTCCACTGAGGCTTCTTCTC RNA extraction and mRNA expression analysis All the gene expression analyses across different experiments were performed in six biological replicates per treatment (n = 6), except for tissue-specific expression analysis where four replicates were used (n = 4). Total RNA was extracted from the tissue samples using the TRIzol reagent (Invitrogen, USA) according to the manufacturer’s protocol. An aliquot (3 µL) of each extracted RNA sample was subjected to agarose gel electrophoresis to verify RNA quality based on the ratio of 28 S to 18 S rRNA bands. Additionally, 2 µL of each RNA sample was used to measure the concentration and purity (OD 260/280 ratio) using a microvolume-spectrophotometric method (Take3 Microvolume Plate, BioTek Eon3, VT, USA). RNA extractions were repeated for samples with OD260/280 ratios outside the range of 1.9–2.0. For gene expression analysis, 1 µg of the total RNA was reverse transcribed into cDNA using the PrimeScript Ist strand cDNA synthesis kit (Takara, U.S.A.) and random hexamer as primer, as per the procedure suggested by the manufacturer. Gene expression analysis of target and reference transcripts was performed in ABI-StepOnePlus™ Real-Time PCR System (Applied Biosystem, USA). The 15 µL PCR reaction mixture consisted of 5 µL of the diluted cDNA (1:50), 0.6 µL each gene specific forward and reverse primers (10 mM); 7.5 µL of 2X SYBR Premix Ex TaqII (Tli RNAase H Plus Takara, U.S.A.), 0.3 uL ROX reference dye and 1 µL of nuclease-free water. The thermal cycling conditions in PCR were 95 °C for 30 s for initial denaturation, followed by a two-step amplification program (5 s at 95 °C and 30 s at 60 °C annealing temperature) repeated 40 cycles, and detection of melting curve at the end of the qPCR run with a temperature gradient of 0.3 °C/10 s from 55 to 94 °C. The mRNA expression assay was performed in technical duplicates for each sample and had three negative controls per assay (no reverse transcriptase, no RNA samples and nuclease free water). A five-point standard curve was used to estimate the assay efficiency in each PCR run, with serially diluted template (15.63–0.98 ng of corresponding RNA). Stability evaluation and validation of reference transcripts Four widely used algorithms – geNorm1, BestKeeper61, NormFinder62, and Delta CT63 were employed to evaluate the stability of the selected reference transcripts (β-act, ef1α, gapdh, and 18 S rRNA) under different experimental conditions. The geNorm analysis was conducted using qbase+ v3.464, whereas the remaining stability assessment algorithms, along with the comprehensive ranking of the most stably expressed reference transcripts, were performed using the RefFinder online tool (https://blooge.cn/RefFinder/)65. The geNorm algorithm is based on the assumption that the ratio of expression value between two good reference genes shall be identical irrespective of the experimental conditions or tissue types. Therefore, geNorm algorithm identifies the most stable combinations of two or more reference transcripts by estimating gene-stability measure (M values) and the pairwise variation between single candidate reference transcript over the other. The candidate transcripts with least ‘M value’ are considered to have the most stable expression pattern. The highly stable reference gene combination has M < 0.5; intermediate M = 0.5–1.0.5.0; least stable M = 1.0–1.5.0.5; and when M > 1.5 candidate reference gene may not be suitable as an internal control1, 64. The geNorm algorithm also provides an additional information on the minimum numbers of reference genes combinations required for satisfactory gene expression analysis by calculating pairwise variations (V values) among the candidate reference transcripts. V = 0.15 is considered as cut-off value, i.e., if the pairwise variation between two or more candidate reference gene is ≤ 0.15, it is not required to consider additional combination of candidate reference transcript and for V > 0.15, additional candidates need to be considered1, 64. NormFinder is a statistical algorithm, which is built on a model-based approach. Unlike the geNorm’s pairwise approach, NormFinder analyses the stability of each candidate reference transcripts independently and ranks them based on the intra and intergroup variation between the subgroup of samples, along with the overall expression variation. The candidate reference genes are ranked in the order of minimum to maximum expression variations61. NormFinder could be particularly useful when there is co-regulation among the candidate reference genes. BestKeeper is a statistical algorithm which identifies expression stability of the candidate reference gene using descriptive statistics such as standard deviation (SD) and coefficient of variation (CV%). The variations in expression of candidate reference gene are negatively correlated to stability and SD > 1 is considered as an unacceptable normaliser gene. BestKeeper also provides weighted expression index (Best Keeper Index, i.e., geometric mean of expression value of stable candidate gene) which can be further used for normalising target gene expression value instead of using one reference gene. BestKeeper also generates a pairwise correlation matrix with Pearson’s correlation coefficient (r) and the corresponding probability (p) values. The correlation matrix between each candidate reference gene and the combined index of the best-correlating genes can be used to assess the relationships between reference and target transcripts62. The delta CT approach is one of the simplest methods for identifying stable reference genes. This method assesses the pairwise variation (ΔCT) in the expression of two candidate reference genes within each sample. When the ΔCT values between the pairs of candidate genes remain consistent, it suggests a stable or coregulated expression pattern. Introducing additional candidate reference genes for pairwise comparisons enables the user to rank or discard them based on their standard deviation. The highly ranked reference gene/s are further selected for normalization of the expression of target genes63. RefFinder is a web-based online tool capable of analysing the stability of reference genes. It utilizes all four algorithms (geNorm, NormFinder, BestKeeper, and delta CT) and combines their results by calculating the geometric mean of the assigned ranks from the four algorithms, producing a comprehensive ranking of candidate reference genes65. For validation of candidate reference transcripts, relative quantification of target genes (myod and myf5) was performed using the mathematical model described by Pfaffl66 after correcting for reaction efficiency. Statistical analysis The transcriptional abundance of myod and myf5 at different rearing temperatures was statistically analysed after normalization with different reference gene(s). Differences among treatments were evaluated using one-way analysis of variance (ANOVA) followed by Tukey’s post hoc multiple comparison test. Statistical significance was indicated by different superscript letters above the bars when p < 0.05. All statistical analyses and graphical visualizations were performed using GraphPad Prism 9. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements RM duly acknowledges the Ph.D. research grant awarded by Asian Fisheries Society through the Kanazawa Research Fellowship. Data for reference gene expression in the early development (Exp.02) were generated as part of RJ’s doctoral research. We sincerely thank Ms. Jettiboina Mahija for providing the samples of the temperature experiment. Abbreviations Open reading frames Coding sequences Untranslated region Deoxynucleic acid Messenger ribonucleic acid Rapid amplification of cDNA ends Polymerase chain reaction Housekeeping genes Beta actin Elongation factor 1 alpha Glyceraldehyde 3-phosphate dehydrogenase 18 S ribosomal RNA Author contributions Manchi Rajesh: Conceptualization, Methodology, Investigation, Software, Formal analysis, Writing-original draft. Biju Sam Kamalam: Conceptualization, Methodology, Investigation Writing – review and editing. Rini Joshi: Investigation and Methodology. Ciji Alexander: Investigation, Writing - review and editing; Prakash Sharma: Investigation, Formal analysis, Md Shabaz Akhtar: Resources and Investigation, All the authors read and approved the final manuscript. Data availability All the data are presented in this article, and raw data are available on reasonable request. Gene sequences were submitted to GenBank sequence repository with accession numbers KU715835, KU715836 and MH616628. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement All the data are presented in this article, and raw data are available on reasonable request. Gene sequences were submitted to GenBank sequence repository with accession numbers KU715835, KU715836 and MH616628.