Novel EST-SNPs Polymorphisms and their Association with Growth Traits in Schizothorax prenanti

1Key Laboratory of Freshwater Fish Reproduction and Development (Ministry of Education), Southwest University, College of Fisheries,, Chongqing 400715, China 2Fisheries Research Institute in Wanzhou Chongqing, Chongqing, 404000, China 3State Key Laboratory of Marine Resources Utilization, South China Sea (Ministry of Education), Hainan University, College of Ocean , Haikou 570100, China Article Information Received 05 June 2019 Revised 20 June 2019 Accepted 10 November 2020 Available online 15 June 2022 (early access)


INTRODUCTION
W ith the development of genomics technology, it is likely that productivity and commercial value could be improved when genomic methods are applied to select superior parents. Marker-assisted selection (MAS) is a powerful method to improve and develop high-quality strains. Compared with traditional methods used in animals, MAS accelerates genetic improvement and the achievement of breeding goals (De-Santis, 2007). It is not affected by the external environment and age stages, which can also use genetic markers for early selection to shorten generational intervals and increase selection intensity (Lu and Wu, 2002).
Single nucleotide polymorphism (SNP) describes polymorphisms caused by point mutations that give rise to different alleles containing alternative bases at a given nucleotide position within a locus (Liu and Cordes, 2004). SNPs have been widely exploited in molecular marker development and genome mapping due to their high abundance, genotyping efficiency, data quality, and genome-wide coverage (Emahazion et al., 1999;Liu and Cordes, 2004;Wang et al., 1998). Moreover, SNPs developed from ESTs are type I markers, which have the advantages of good versatility, clear bands, and analytical simplicity. SNPs from candidate genes are becoming important and efficient molecular markers for MAS (Spelman et al., 1999). In recent years, SNP markers associated with important traits have been reported in animal and plant, especially growth trait. In aquatic animals, SNP markers associated with growth traits have been reported in Chlamys farreri (Guo et al., 2012), Pelteobagrus fulvidraco , Crassostrea gigas (Cong et al., 2013;Cong et al., 2014), Micrpterus salmoides , Macrobrachium rosenbergii (Thanh et al., 2010), Siniperca chuatsiand (Dong et al., 2019) and so on. Schizothorax prenanti, one of the indigenous economic fish species in China, mainly distributed in the upper reaches of the Yangtze River and its tributaries with O n l i n e

F i r s t A r t i c l e
rich nutritional value is of high economic value (Chen and Cao, 2000;Wu and Wu, 1992). With the increase in artificial breeding intensification, the germplasm resources of S. prenanti have been degraded, which is revealed in individual miniaturization, slow growth, and decreased disease resistance Ye et al., 2018). Therefore, there is an obvious need for more molecular markers associated with growth traits for intensive study on the molecular marker-assisted breeding. Although the SNP of S. prenanti have been developed in recent years , the genetic molecular markers associated with growth traits remain scarce. In this study, 31 novel EST-SNP markers were analyzed their polymorphism using Mass ARRAY ® MALDI-TOF System. To provide valuable information in molecular assisted breeding and genetic improvement for S. prenanti, we explored associations between SNP markers and growth traits in the cultured populations of S. prenanti.

Sample collection, genomic DNA preparation and commercial traits measurement
The experimental protocols were approved by the institutional animal care and use committee of Southwest University. A total of 164 individuals were randomly selected from a cultured population in E Meishan, (Sichuan Province, China) in July 2017. The samples were bred in the same batch and pond. Fishes were anesthetized using tricaine meth-anesulfonate (MS222). Then Total length (TL), body length (BL) and body height (BH) were measured by Vernier Caliper (0.1 mm accuracy). Body weight (BW) was weighed using an electronic balance (0.1 g accuracy). After obtaining the measurement data, fins were cut and preserved in 95% alcohol, and genomic DNAs were extracted using Animal genomic DNA extraction kit (Sangon biotech, Shang Hai, China). DNA was examined by performing 1% agarose gel electrophoresis, and concentrations were determined using photometry (Eppendorf, German). The working DNA concentration was 30 ng/uL.

Polymorphism SNP loci screening and SNP genotyping
EST-SNP sequences were obtained from transcriptome of Spleen in S. prenanti in our previous research . Based on EST-SNP sequence information, primers were designed (Table I). PCR was performed in a 10 uL reaction volume, containing 1.25 uL 10×PCR Buffer with 15 mM MgCl 2 , 0.65 uL 25 mM MgCl 2 , 2 uL dNTP (2.5 mM), 2 uL forward primer (0.5 uM), 2 uL reverse primer (0.5 uM), 0.2 uL ExTaq DNA polymerase (5 U/ uL) (TaKaRa), using 2 uL genome DNA as template, adding 1.9 uL ddH 2 O. PCR profiles included an initial denaturation at 94 °C for 15 min, 94°C denaturation for 20 s, annealing temperature 56 °C for 30 s, 72 °C for 60 s for 45 cycles and a final extension at 72 °C for 3 min. SNP sequence-specific extension primers were added to the PCR-amplified product and a base was extended at the SNP site. The extended product was purified and cocrystallized with a MassARRAY® SpectroCHIP chip with a surface-covered substrate, and the crystals were placed in a vacuum tube of a mass spectrometer to automatically analyze the SNPs site information.

Data analysis
Observed heterozygosity (Ho), expected heterozygosity (He) and Hardy-Weinberg equilibrium (HWE) for SNP loci were calculated using POPGENE version 32 software (Yeh and Boyle, 2000). Polymorphic information content (PIC) was calculated by PIC-CALC software.
Morphological analysis was performed with SPSS 19.0. Multivariate analysis of variance and independentsamples T test in the general linear model (GLM) were used to analyze the association between SNP locus and growth traits. Significant differences were tested using Duncan's multiple range test. Differences were considered to be statistically significant when p < 0.05 or p<0.01.

Morphological analysis
The descriptive statistics including mean, skewness, kurtosis, minimum and maximum for the growth traits were summarized in Table II. Skewness and kurtosis are parameters to test normal distribution. Skewness test and kurtosis test both indicated the four traits (BW, TL, BL and BH) obeyed normal distribution (p>0.05) (West et al., 1995).
Pearson correlation analysis showed that there was a significant correlation between the traits (p<0.01) (Table  III). Maximum correlation coefficient was 0.982 between BL and TL. Minimum correlation coefficient was 0.887 between TL and BH.
Principal component analysis (PCA) indicated that body weight accounting for 94.50% of the variance, the eigenvalue more than 1, the accumulative variance more than 85%, it was the first principal component of the growth traits of S. prenanti (Table IV).

The characters of SNP loci
Point mutations of SNP included transitions and transversions. Among the 31 SNP markers, the number of transitions and transversions was respectively 18 and 13, and the ratio between them was approximately 1.38.

O n l i n e F i r s t A r t i c l e
23 loci were polymorphic and the proportion was 74.2%. Genetic parameters of 23 SNPs are shown in Table V (Botstein et al., 1980). While the remained 10 loci had low polymorphism (PIC<0.25). 13 of the 23 SNPs were in accordance with HWE (p>0.05).

Associations between EST-SNPs and growth traits
The results of multivariate analysis of variance in the general linear model indicated that CC genotype at ug25066-1502 had significantly higher values for BW and TL than did individuals with the GG genotype (p<0.05), and CC genotype had significantly greater values for BW than CG genotype (p<0.05). For ug23056-2976, TT genotype was significantly lower than CT genotype for BH (P<0.05). Moreover, compared with CC genotype, individuals with the AC genotype at ug22539-1605 had significantly higher value in TL, BL, BH, and BW (p<0.05) (Table VI). Further analysed, the allele C was significantly higher than allele A in TL, BL and BH (Table VII) (p<0.05).

DISCUSSION
Principal component analysis is a dimension reduction technique that used to describe the relations between several response variables and explain the total variation in the data (Abbas and Wasin., 2019). To date, PCA has been widely used in aquatic area, such as aquatic ecosystem (Uddameri et al., 2014), aquatic nutrition (Casu et al., 2017;Gammanpila et al., 2017), morphological analysis (Jiang et al., 2012;Li et al., 2015). It is useful when the variables under study are highly correlated. In our study, the correlation analysis showed that four traits (BW, TL, BL and BH) have an extremely significant relationship. What is more, the result of PCA showed that BW is the first principal component of the growth traits of S. prenanti. Therefore, body weight is a main indicator in selective breeding of S. prenanti. Furthermore, TL, BH and BL could be indirect indicators to reflect the growth.  Abundant SNPs discovered by the next generation sequencing technologies have allowed us to better understand the association between genomic variation and production traits in aquatic species (Yáñez et al., 2014). In this study, we identified 31 genomic SNP loci from the unigene data of the transcriptome. The ratio between conversion and transversion was 1.38, similar to turbot (1.35) (Vera et al., 2013), lower than lake sturgeon (1.65) (Hale et al., 2009), which is related to the differences of the species' own genome and living environment. The ratio between conversion and transversion in point mutations has a great influence on the degree of gene selection pressure and reflected deviation of mutation. Besides, the data of genotyping of SNP in 164 S. prenanti showed 23 loci were polymorphic accounting for 74.2%, which was lower than previous research in S. prenanti  and higher than Mauremys mutica (Zhao et al., 2016) and large yellow croaker (Jiang et al., 2015).

Novel EST-SNPs Polymorphisms and their Association with Growth Traits
Heterozygosity could measure the genetic variability; He>5 indicated that the population experiences low selection and maintains higher genetic diversity (Jiang et al., 2015;Zhang et al., 2019). Our data suggested that the mean value of observed heterozygosity and expected heterozygosity were respectively 0.2094 and 0.2019, which indicated the population has low genetic diversity. After calculating the PIC in the population, 13 loci were found to have moderate polymorphism (0.25≤PIC<0.5), whereas others were low polymorphism (PIC<0.25). PIC demonstrates the degree of DNA variation, the value of which is dependent on the number of alleles and their frequency distribution. If the number of alleles was greater and the allelic frequencies of all alleles were more balanced, the PIC will be greater (nearly 1) . Low genetic parameters could be attributed to the fact that the SNP only has two alleles, allelic imbalance, and the samples from cultured population and the population's genetic diversity were low.
If there is a significant association between markers and specific trait in a population, correlation analysis can figure out that which marker is associated with that trait (Doerge, 2002;Lynch and Walsh, 1997). This association has already reached the significant level, which may suggest the relationships between the marker and the trait. Consequently, the selective breeding inferred from phenotype can operate to genotype-assisted selection. In this study, we found that CC genotype at ug25066-1502 was significantly superior than GC genotype and GG genotype with higher BH, which demonstrates that ug25066-1502 was correlated with BH. However, only BL in the CC genotype was significantly superior than the GG genotype, while no significant difference was observed between GG and GC genotype. Therefore, these data imply that ug23056-2976 may have some effects on TL, but the effects are not major. Apart from these, at ug23056-2976 and ug22539-1605, we only detected two genotypes. The results imply that the CC at ug23056-2976 and AA at ug22539-1605 are rare genotypes. Or we need additional work to utilize larger samples for further confirmation. For ug22539-1605, the CC genotype was significantly correlated with the four traits. In our analysis, we found O n l i n e

F i r s t A r t i c l e
Y. Yang et al. that the allelic genotype C was significantly higher than allele A in TL, BL and BH. So, the allelic genotype C may be had a major impact on growth traits.

CONCLUSIONS
In our study, principal component analysis indicated that body weight was the first principal component of the growth traits of S. prenanti. And we found that ug25066-1502 was correlated with BH and ug22539-1605 was significantly correlated with the four traits. The two loci could be used as important candidate molecular markers for selective breeding of S. prenanti.

O n l i n e F i r s t A r t i c l e
Novel EST-SNPs Polymorphisms and their Association with Growth Traits