Genome-wide association study of molecular markers related to keel length in a specialized strain of yellow feathered broilers

Abstract

Keel length is a key body size indicator, which has an important impact on the overall growth performance, bone health, and production performance of poultry. In the process of selecting crossed strains for yellow feathered broiler chickens, it is generally necessary to select keel length, but there is little research on keel length development. Therefore, resequencing and GWAS was employed to obtain SNP molecular markers associated with keel length in a specialized Yellow-feathered broiler line. We identified 10 SNPs that were potentially significantly correlated with keel length for the first time located at 9 genes, including ATP7B, CST3, OTOP1, CRMP1, SLC12A1, COPS2, FAM227B, IFT140, APLP2. SNP2 and SNP3 loci were in a strongly linked state (D ' value=1), the other 8 SNP loci were not in a strongly linked state (D 'value<1). The association analysis between single SNP marker and keel length traits showed that TT and CT at SNP1 (rs315701680), GG at SNP2 (rs738740137), AA at SNP3 (rs317223723), AA at SNP4 (rs732443622), GG at SNP5(rs315667756), CC at SNP6 (rs314381113), GG at SNP7 (rs732811384), TT at SNP8 (rs314197610), CC and TC at SNP9 (rs13782000), TT and GT at SNP10 (rs737401141) genotypes were all the dominant genotypes for keel length. The strong linkage between SNP2 and SNP3 resulted in two haplotypes H1 (AG) and H2 (GA), respectively. The H1H1 haplotype (GGAA) produced by SNP2 and SNP3 linkage was the dominant genotype for keel length. The SNP molecular markers and dominant genotypes at SNP1-SNP10 loci identified in this study may be used to improve the accuracy and genetic progress of keel length selection. Meanwhile, candidate genes potentially significantly related to keel length will lay the foundation for genetic selection of keel length and cultivation of high-quality yellow feathered broilers in the future.

Introduction

Due to long-term natural and/or artificial selection, different chicken breeds exhibit significant differences in body size (Lyu et al., 2023). The body size of chickens reflects their physical development, particularly the growth of bones and muscles (Gao et al., 2011; Geng et al., 2021). Key body size traits such as keel length, body slope length, chest width, and chest depth serve as important indicators of body size. Among these, keel length is a critical body size trait during poultry growth, reflecting the development of the skeletal system and significantly influencing overall growth performance, skeletal health, breast muscle percentage, and egg-laying performance (Duenk et al., 2025; Luo et al., 2024). There is a strong positive correlation between keel length and breast muscle thickness, as breast muscle primarily attaches to the sternum and keel, closely linked to keel growth. A significant correlation exists between keel length and body weight of chickens, allowing for improved body weight and breast muscle percentage through selective breeding for keel length (Li et al., 2021; Lyu et al., 2023). In the breeding process of yellow-feathered broilers, specialized strains (especially terminal sire strains) are selected for body weight, breast muscle percentage, and body size, in order to produce commercial broilers with faster growth rates, higher muscle percentages, and better body uniformity.Currently, the body size uniformity of yellow-feathered broilers is poor, and body size selection in breeding primarily relies on simple measurements of body dimensions, resulting in slow genetic progress of body size uniformity. It is of great significance to identify candidate genes and molecular markers related to keel length for the application of body type breeding, in order to accelerate the progress of body weight, breast muscle percentage, and body shape uniformity breeding. With the emergence of high-density SNP sequencing and genome-wide association studies (GWAS) technology, SNP molecular markers and important candidate genes in keel length can be more accurately detected. Based on 580 K chips and GWAS technology, Lyu et al. identified QTLs affecting chicken keel length and weight in the overlapping region between 10926588 and 10933626 bp on GGA6 in four larger-sized Asian fighting chickens and five smaller-sized chickens (Lyu et al., 2023). GH, ITM2B, CAB39L, SETDB2, PHF11, ARL11, KPNA3, SPRYD7, SLIT2, LCORL, and LDB2 were important candidate genes influencing the keel length of Wenshang Barred chickens at 18 weeks of age (Wang et al., 2024; Yang et al., 2022). The GG genotype at the PITX2 gene g.13726A>G locus was the dominant genotype for keel length in Wuliangshan black bone chickens (Cao et al., 2019). The SSC13 haplotype was significantly correlated with keel length in Yandang Partridge chickens (Wu et al., 2023). Based on multiple previous studies, candidate genes related to keel length vary among different chicken breeds and populations, indicating that keel length is a complex quantitative trait controlled by multiple genes and signaling pathways. There is relatively little research on candidate genes and molecular markers related to keel length in specialized strains of yellow feathered broiler chickens. Herein, we used resequencing and GWAS to obtain SNP molecular markers associated with keel length in a specialized strain of Yellow-feathered broiler. We expect that the SNP molecular markers identified in this study can be used to improve the accuracy and genetic progress of keel length selection. Our findings will lay the foundation for genetic selection of keel length and more effectively cultivation of high-quality yellow feathered broilers with longer keel length, larger body weight, and better breast muscle percentage in the future.

Material and methods

Chicken population

Four hundred males randomly selected from the 12nd generations terminal strain of fast-type yellow feathered broiler in Jiangsu Lihua Breeding Co., Ltd. This strain is involved in the formation of commercial chickens through hybridization, with a body weight of about 2.5 kg after 60 days. All chickens were produced under the same condition and immunization follows the standard procedures established by the company. During the experimental period, the chicken flock was free access to water.

Measurement of keel length

The keel length is measured by measuring the distance from the front end of the keel bulge to the end of the keel along the body surface with a caliper along a straight line at 60 days. The detailed measurement methods are as follows: one person hold a chicken, one hand holds 2 legs, the other hand holds one wing, and the chicken naturally lies flat on the table with its chest facing up. The other person holds a vernier caliper in their right hand, and places their thumb on the front end of the keel bulge with their left hand, and their index or middle finger on the end of the keel. The distance measured by the vernier caliper along the straight line from the front end of the keel bulge and end of the keel held by their left hand is the keel length. Details of keel length measurement was seen in Fig. 1. The keel length of each chicken was measured once, and finally take the average of all individuals to obtain the keel length of the population. All measurements of keel length were taken by the same person, and the measurement locations were basically the same.

Fig. 1.

Details of diagrammatic methods of keel length measurement.

Whole-genome resequencing and quality control

1.5 mL blood were obtained from the vein of the wings of all chicken populations whose keel length had been measured at 60 d, and DNA was extracted using the conventional phenol-chloroform method. The whole genome resequencing of experimental chicken DNA samples was completed by Shenzhen Huada Gene Technology Co., Ltd. (China) with a sequencing depth of 10 × . The raw reads from the sequenced genome were filtered using the fastp (v.0.20.0) preprocessor to remove low-quality reads, adapters, and reads containing poly-N sequences (Bush, 2009). The quality-controlled reads were then aligned to the reference genome GRcg7b using the BWA (v0.7.17) software (Li and Durbin, 2009). The reads were converted in format and sorted by Picard 3.1 and SAMtools 1.18, with duplicate alignments removed. SNPs were detected using GATK 4.0 software, yielding high-quality SNP loci, totaling 16,411,878 SNP loci detected (Niaré et al., 2023). The quality control standards for loci and individuals are as follows. SNP site deletion frequency was lower 2%, allele frequency maf was more than 0.05, Hardy Weinberg P-value <1 × 10^-4, while excluding individuals genotype deletion rates higher than 4% (Tu et al., 2025). Finally, 366 individuals and 10266594 SNP loci were obtained for further Genome wide association study (GWAS).

Genome-wide association study

GWAS analysis of keel length traits was conducted based on the Linear Mixed Model (LMM) (Fan et al., 2025). The LMM is as follows.

$$\mathrm{y}=\mathrm{W}\alpha +\mathrm{x}\beta +\mathrm{u}+\in \mathrm{u}\sim \text{MVNn}\left(0,\lambda {\tau }^{-1}\mathrm{K}\right),\in \sim \text{MVNn}\left(0,\lambda {\tau }^{-1}\text{In}\right)$$

Where y is the phenotype value for all individuals, W is a covariance matrix used to control the population structure (the first three principal components including PCA), α is a fixed effect vector, X is the genotype of the SNP, β is the effect size of the SNP for the phenotypes, u is a vector of random effects with a covariance structure that refers to the n-dimensional multivariate normal distribution u∼ MVNn (0, λτ⁻¹K). and ϵ represents the residual vector. MVNn is an n-dimensional multivariate normal distribution, λ is the ratio of genetic variance to residual variance (random effect vector u and residual vector ϵ), τ⁻¹ is the residual variance, K is the kinship matrix calculated based on SNPs using GEMMA, and In is the identity matrix.

Using PLINK software, 432,529 independent SNP markers were obtained by trimming linkage disequilibrium with parameters set to 50 SNP window sizes, 10 SNP step sizes, and an r² threshold of 0.2. After multiple tests, a 5% Bonferroni correction method was used to set the significance threshold. Based on this, the genome-wide significance threshold and potential significance threshold were calculated as P = 1.16 × 10^-7 (0.05/432,529), and P = 2.31 × 10^-6 (1/432,529), respectively. Thus, the 5% Bonferroni genome-wide significance or potential significance threshold was -log10 ^(P-value) =6.94 or -log10^(P-value)=5.64, respectively. The CMplot package in R language was used to generate Manhattan plots and quantile quantile (Q-Q) plots (Yin et al., 2021).

Statistical analysis

The association analysis between genotype and keel length was done based on ANOVA in the general linear model of SPSS 20.0. Among them, the different genotypes were fixed factors, and keel length was the dependent variable. Tukey HSD method was used to compare the significant differences in keel length between different marker genotypes, and identify the dominant genotype. Linkage disequilibrium (LD) analysis on SNP loci was done using Hapoview4.1, and obtained genetic diversity indicators such as genotype frequency, gene frequency, and heterozygosity (He) for 10 SNP loci. Chi square (χ²) test was used to detect whether SNP loci are in H-W equilibrium state.

Results

Keel length characteristics analysis

At 60 days of age, the average keel length of 366 experimental chickens was 106.44 mm, with the maximum and minimum values of 127.17 mm and 89.38 mm, respectively, and a coefficient of variation of 6.11%, indicating that the genetic variation of keel length in the population was relatively small.

Genome-wide association study of keel length

The Manhattan and Q-Q plots for keel length are shown in Fig. 2. and Fig. 3., respectively. According to Fig. 2. and Table 1, 10 SNPs potentially significantly associated (P < 2.31 × 10^-6) with keel length were identified on chromosomes 1, 3, 4, 10, 14, and 24. These loci are located at introns, upstream or downstream of a total of 9 genes, including ATP7B, CST3, OTOP1, CRMP1, SLC12A1, COPS2, FAM227B, IFT140, and APLP2.

Fig. 2.

Manhattan plots for genome-wide association study(GWAS) of keel length. The upper red dotted line represents the 5% Bonferroni genome-wide significance threshold [-log10 (P-value) =6.94], and the lower red dotted line represents the potential significance threshold[-log10(P-value)=5.64]. The highlighted area represents candidate genes or the nearest genes within 100 kb upstream and downstream for significant SNP loci.

Fig. 3.

Quantile–quantile (Q-Q)plots for genome-wide association study(GWAS) of keel length. The inflation coefficients (λ) for keel length is 1.03.

Table 1.

Annotations of 5% bonferroni genome-wide suggestive-level SNPs associated with keel length.

SNP loci	SNP ID1	Position2	Ref/Alt3	P Value4	Candidate/Surrounding gene (100 kb)5	Location
SNP1	rs315701680	1:170194700	C/T	1.73 × 10-7	ATP7B	upstream
SNP2	rs738740137	3:15305633	G/A	2.21 × 10-6	CST3	downstream
SNP3	rs317223723	3:15305745	A/G	9.94 × 10-7	CST3	downstream
SNP4	rs732443622	4:78036233	G/A	2.05 × 10-6	OTOP1	intron
SNP5	rs315667756	4:78750083	A/G	1.36 × 10-6	CRMP1	intron
SNP6	rs314381113	10:10111788	C/A	2.22 × 10-6	SLC12A1	intron
SNP7	rs732811384	10:10468445	A/G	1.21 × 10-6	COPS2	intron
SNP8	rs314197610	10:10534417	G/T	1.02 × 10-6	FAM227B	intron
SNP9	rs13782000	14:13797770	T/C	2.14 × 10-6	IFT140	intron
SNP10	rs737401141	24:1536829	G/T	1.50 × 10-6	APLP2	downstream

¹SNP: rs ID of the locus.

²Position: Physical location of the SNP (chromosome: position).

³Alt/Ref: Alternative allele and reference allele.

⁴P value:Hardy Weinberg frequency P-value.

⁵Candidate/Surrounding gene (100 kb): Genes annotated in the 100 kb region upstream and downstream of the SNP.

The Q-Q plots results are shown in Fig. 3. The expansion coefficient (λ) of the keel length was 1.03 (λ ≈ 1) , indicating that the model selection was reasonable, and strong correlation sites with the target trait have been screened, and the analysis results was reliable (Georgiopoulos and Evangelou, 2016; Guintivano et al., 2020).

Genetic characteristics analysis of 10 SNP loci

According to the H-W balance test, SNP5, SNP6, SNP10 locus deviating from the H-W balance state (P < 0.05) (Table 2). The conditions for the H-W equilibrium hypothesis included a sufficiently large population, random mating, absence of mutations and migration, and no genotype selection advantage (Zou and Donner, 2006). We found that several SNP loci deviated from the H-W equilibrium, which possibly due to artificial mating, selection pressure of body weight and FCR for this specialized strain of yellow feathered broilers. Except for SNP1, SNP2, SNP3, SNP10 locus with low genetic diversity (0.10 < He < 0.25), the other 6 SNP loci belonged to moderate genetic diversity (0.25 < He <0.50), which can provide more genetic variation among individuals, construct heterotic groups with longer keel length(Atzmon et al., 2002; Duenk et al., 2021).

Table 2.

The genetic characteristics of 10 SNP loci.

SNP loci	SNP ID	Genetype	Frequency of gene type	Allele	Allele frequency	He	Ho	H-W (P value)
SNP1	rs315701680	C C	0.82	C	0.91	0.165	0.169	0.769
		C T	0.16	T	0.09
		T T	0.01
SNP2	rs738740137	A A	0.01	A	0.10	0.175	0.177	0.946
		G A	0.17	G	0.90
		G G	0.81
SNP3	rs317223723	A A	0.81	A	0.89	0.169	0.19	0.076
		A G	0.17	G	0.11
		G G	0.02
SNP4	rs732443622	A A	0.10	A	0.24	0.292	0.308	0.413
		G A	0.29	G	0.76
		G G	0.61
SNP5	rs315667756	A A	0.62	A	0.77	0.31	0.353	0.029
		A G	0.31	G	0.23
		G G	0.07
SNP6	rs314381113	A A	0.46	A	0.66	0.396	0.449	0.032
		C A	0.40	C	0.34
		C C	0.14
SNP7	rs732811384	A A	0.44	A	0.34	0.445	0.449	0.948
		A G	0.45	G	0.66
		G G	0.12
SNP8	rs314197610	G G	0.39	G	0.61	0.433	0.477	0.095
		G T	0.43	T	0.39
		T T	0.18
SNP9	rs13782000	C C	0.16	C	0.39	0.463	0.475	0.696
		T C	0.46	T	0.61
		T T	0.38
SNP10	rs737401141	G G	0.82	G	0.90	0.152	0.181	0.010
		G T	0.15	T	0.10
		T T	0.02

Dominant genotypes analysis of SNPs loci for keel length

Firstly, we analyzed the correlation between a single SNP marker and keel length, as shown in Table 3. The keel length of individuals with TT and CT genotypes at SNP1 locus (rs315701680) was significantly longer than that of individuals with CC genotype (P < 0.05). The keel length of individuals with GG genotypes at SNP2 (rs738740137) locus was significantly longer than that of individuals with AA and GA genotype (P < 0.05). The keel length of individuals with AA genotypes at SNP3 (rs317223723) locus was significantly greater than that of individuals with GG and AG genotype (P < 0.05). The keel length of individuals with AA genotypes at SNP4 (rs732443622) locus was significantly longer than that of individuals with GG and GA genotype (P < 0.05). The keel length of individuals with GG genotypes at SNP5 (rs315667756) was significantly longer than that of individuals with AA and GA genotype (P < 0.05). The keel length of individuals with CC genotype at SNP6 (rs314381113) was significantly longer than that of individuals with AA and CA genotypes (P < 0.05). The keel length of individuals with SNP7 (rs732811384) genotype GG was significantly longer than that of individuals with AA and AG genotypes (P < 0.05). The keel length of individuals with SNP8 (rs314197610) locus TT genotypes was significantly longer than that of individuals with GG and GT genotype (P < 0.05).The keel length of individuals with CC and TC genotype at SNP9 locus (rs13782000) was significantly longer than that of individuals with TT genotypes (P < 0.05). The keel length of individuals with SNP10 (rs737401141) genotype TT and GT was significantly longer than that of individuals with GG genotypes (P < 0.05).

Table 3.

Association analysis between 10 SNP loci and keel length (mean ± SD).

SNP loci	Gene type	Keel length (mm)	Number of gene type
SNP1	C C	105.64±6.27b	300
	C T	109.92±6.48a	60
	T T	112.11±6.29a	4
SNP2	A A	103.19±8.36b	4
	G A	103.03±5.83b	64
	G G	107.21±6.39a	298
SNP3	A A	107.25±6.39a	296
	A G	103.08±5.75b	62
	G G	102.27±6.97b	8
SNP4	A A	112.22±7.23a	36
	G A	108.12±6.50b	107
	G G	105.32±6.11c	223
SNP5	A A	105.32±6.23c	225
	A G	107.46±6.31b	113
	G G	111.74±6.43a	27
SNP6	A A	104.82±6.00c	169
	A C	107.22±6.62b	145
	C C	109.51±6.35a	52
SNP7	A A	104.67±6.22c	160
	A G	107.37±6.39b	163
	G G	109.51±6.23a	43
SNP8	G G	104.68±6.22c	143
	G T	106.85±6.36b	158
	T T	109.29±6.42a	64
SNP9	C C	108.71±6.05a	57
	T C	107.48±6.09a	169
	T T	104.27±6.60b	139
SNP10	G G	105.78±6.27b	299
	G T	109.46±6.80a	55
	T T	110.78±6.99a	9

Note: Different lowercase letters indicate that the different genotypes at the same locus are significantly different (P<0.05).

Detailed results of the linkage disequilibrium analysis(LD) are shown in Fig. 4. According to Fig. 4A. and Fig. 4B., SNP2 and SNP3 loci are in a strongly linked state (D 'value=1), the other 8 SNP loci are not in a strongly linked state(D 'value<1) . According to Fig. 4C.,The strong linkage between SNP2 and SNP3 resulted in two haplotypes H1 (AG) and H2 (GA), with haplotype frequency of 0.893 and 0.098, respectively. The combination of two haplotypes produced by SNP2 and SNP3 linkage resulted in five genotypes (Table 4). As shown in Table 4, association analysis revealed that the individuals with H1H1 haplotype (GGAA) had longer keel length than those of the of the individuals with other haplotypes combinations.

Fig. 4.

The linkage disequilibrium analysis among 10 SNP loci for keel length (A) Linkage disequilibrium plot of SNP1-SNP10. (B) Linkage disequilibrium plot of SNP2-SNP3. (C) The haplotype and haplotype frenquency produced after SNP2-SNP3 linkage. The color of the square ranges from light to dark indicates that the degree of linkage from low to high. The number in the square represents the strength of the linkage disequilibrium between the sites (number = D, value × 100).

Table 4.

Association analysis between SNP2 and SNP3 haplotype combinations and keel length (mean ± SD).

Haplotype combination	Gene type	Keel length (mm)	Number of gene type
H2H2	AAGG	103.19±8.36b	4
H1H2	GAAG	103.14±5.84b	60
H2H1	GAGG	101.35±6.41b	4
H1H1	GGAA	107.25±6.39a	296
H2H1	GGAG	101.32±1.67b	2

Note: Different lowercase letters in the same column indicate significant differences between different genotypes of keel length (P<0.05).

Discussion

Keel length is a commonly indicator for evaluating the growth and development of poultry bones, and is also one of the important body size selection traits in poultry breeding. With high-intensity breeding of body weight and egg production, the production performance of chickens has almost approached its physiological limit, and the economic losses caused by some metabolic diseases have increasingly attracted the attention of farmers. Among them, bone diseases are particularly evident, such as keel bone fractures, rickets in broilers (Dinev, 2012; Duenk et al., 2025). These bone diseases are mainly caused by a mismatch between high levels of productivity and low levels of bone development. Therefore, study keel length development is considered an effective way of solving bone metabolic diseases. Keel length is a skeletal trait, indicating that this genome region may contain genes that affect bone development and calcium homeostasis.

Through resequencing and GWAS, we identified 10 SNP loci potentially significantly associated with keel length in a specialized yellow-feathered broilers, along with several important candidate genes. These findings differed from previously discovered SNP molecular markers and candidate genes for keel length in Wenshang barred chicken, Wuliangshan black-bone chickens, and Yandang Partridge chickens (Cao et al., 2019; Wang et al., 2024; Wu et al., 2023; Yang et al., 2022). This discrepancy may be attributed to the fact that keel length is a complex quantitative trait determined by multiple genes and signaling pathways, with varying genetic mechanisms across different chicken breeds and chicken populations at different ages. Some candidate genes such as CST3, ATP7B, OTOP1, CRMP1, COPS2, IFT140 related to keel length in this study are mainly involved in osteoblast differentiation, mineralization, calcium binding, bone formation.

Cystatin C (CST3), also known as cysteine protease inhibitor 3, functions as a protein that inhibits cysteine protease, participates in the regulation of intracellular and extracellular protein hydrolysis, and protects cells from inappropriate endogenous or exogenous protease hydrolysis. CST3 protein is functionally associated with immune response (Biasizzo et al., 2021). In vitro cell experiments have shown that CST3 can promote osteoblast differentiation, mineralization, and bone formation by affecting the BMP signaling cascade in osteoblasts (Danjo et al., 2007; Weivoda et al., 2020), regulate bone homeostasis, and have a protective effect on osteoporosis and related fractures (Ensrud et al., 2013; Zhou et al., 2024). We found that the dominant genotypes of two SNP sites (SNP2 rs738740137 G→A and SNP3 rs317223723 A→G) were GG and AA, respectively. After strong linkage between these two SNP sites, the GGAA haplotype also became the dominant genotype of keel length. Although the two mutations did not increase the keel length, both these two SNP sites were located at downstream of the CST3 gene, indicating that they may be regulated by the CST3 gene, which affects the keel development.

Copper transport P-type adenosine triphosphatase (ATP7B), is a pathogenic gene in human Wilson's disease. It encodes a copper transport P-type ATPase protein that plays a key role in copper metabolism by hydrolyzing ATP to drive copper ion transport (Singh et al., 2019). Chickens can easily absorb and accumulate copper from the environment and food, and the function of ATP7B was to transport excess copper out of the cytoplasm, preventing poultry hepatotoxic death caused by excessive copper deposition (Schushan et al., 2012; Zhong et al., 2023). ATP7B was correlated with the growth rate and meat production performance indicators of broilers (Abdalhag et al., 2015), as well as with the feed consumption and egg laying interval of laying hens (Volkova et al., 2025). TT and CT of SNP1(rs315701680, C→T) locus was the dominant genotypes for keel length, speculating that they may be regulated by it’s downstream gene ATP7B, which may indirectly affect keel development through biological processes such as ATP and calcium ion binding.

The protein encoded by the otopetrin-1 (OTOP1) gene is a proton channel that responds to protons, has high permeability and selectivity towards H⁺, increases intracellular acidity, and produces an uncomfortable response to acid stimulation (Saotome et al., 2019; Zhang et al., 2025). OTOP1 can regulate intracellular Ca²⁺, affect the biomineralization of otoliths (Hughes et al., 2007), participate in maintaining balance of the vestibular system, perceive gravity and limb orientation (Hurle et al., 2003). The loss of OTOP1 function mutation may lead to the influx of calcium ions and disruption of calcium homeostasis, which may result in peg-shaped teeth (Choi et al., 2025). The SNP locus of OTOP1 gene (SNP4, rs732443622 G→A) had a dominant genotype of AA for keel length, indicating that this gene mutation was beneficial and may affect keel development by regulating intracellular Ca²⁺.

Collapsin response mediator protein 1 (CRMP1) primarily participates in neural developmental processes, such as neuronal migration, axonal growth, dendritic branching, and apoptosis mediated by extracellular signaling molecules (e.g., Sema3A, reelin, neurotrophic factors) (Makihara et al., 2016; Su et al., 2007; Yamashita et al., 2006). Phosphorylated CRMP1 exhibits a strong inhibitory effect on the growth of dorsal root ganglion neuron processes and induces their collapse (Quach et al., 2004). Knockout of the CRMP1 gene in mice results in reduced proliferation and apoptosis levels of granule cells in the cerebellum, along with impaired neuronal migration (Charrier et al., 2006). The CRMP1 gene plays a role in growth cone remodeling during neuronal development and regeneration, with higher expression levels of CRMP1 mRNA observed in the peripheral blood and bone marrow of patients with non-high-risk neuroblastoma (Nakatani et al., 2024). The SNP locus (SNP5, rs315667756, A→G) GG genotype of CRMP1 gene was the dominant genotype for keel length, indicating that this gene mutation may affect keel development through physiological processes such as neuronal migration and axonal growth.

COP9 signalome complex subunit 2 (COPS2) is associated with a series of important physiological processes such as cell proliferation, differentiation, and apoptosis (Li et al., 2018; Zhang et al., 2016). Overexpression of COPS2 may be involved in maintaining the proliferation of chicken embryonic fibroblast (CEF) cells (Taye et al., 2017), and its regulated signaling pathway is associated with dysregulation of bone lipid metabolism, making it a key gene in women with osteoporosis (Azuma et al., 2009; Qian et al., 2019).We found that the SNP locus (SNP7, rs732811384, A→G) GG of the COPS2 gene was the dominant genotype for keel length, indicating that this gene mutation may affect keel development through chicken embryonic fibroblast proliferation, differentiation, or related signaling pathways.

Family with sequence similarity 227 member B (FAM227B) may play an important role in the pathogenesis of congenital pseudarthrosis of the tibia (CPT) in humans, closely related to the decreased osteogenic differentiation ability of CPT periosteal stem cells (PSCs) (Li et al., 2023). The SNP locus (SNP8, rs314197610 G→T) TT of FAM227B gene in this study was the dominant genotype for keel length, indicating that this gene mutation is a beneficial mutation for keel length.

Intraflagellar transport 140 (IFT140) mRNA are highly expressed in bones and play important roles in cartilage formation and osteogenic differentiation (Zhang et al., 2019). The IFT140 mutation is associated with skeletal fibrosis (Khan et al., 2014; Miller et al., 2013). We found that the IFT140 gene mutation (SNP9, rs13782000 T→C) CC was the dominant genotype for keel length, indicating that this SNP site is a beneficial mutation for keel length and may play an important role in keel development through processes such as cartilage formation and osteogenic differentiation.

Solute carrier family 12 member 1 (SLC12A1) can transport Na⁺, K⁺, and Cl^- from the extracellular space to the intracellular space (Ares et al., 2011; Kemter et al., 2014;). The SLC12A1 gene encodes the sodium potassium chloride cotransporter NKCC2, which belongs to the sodium dependent subgroup of the SLC12 family. It mediates the electrically neutral transport of sodium and potassium ions, affecting salt reabsorption, acid-base balance regulation, and divalent mineral cation metabolism (Ares et al., 2011; Markadieu et al., 2014). SLC12A1 gene functional loss mutations can cause Bartter syndrome in human kidney disease, characterized by hypokalemia, hypochloremic metabolic alkalosis, and osteopenia (Yang et al., 2024; Yi et al., 2022). We found that the SLC12A1 SNP locus (SNP6, rs314381113, C→A) was correlated with keel length for the first time. This SNP locus may participate in keel development through physiological processes such as salt reabsorption, acid-base balance regulation, and divalent mineral cation metabolism.

In this study, 10 SNPs were identified to be potentially significantly correlated with keel length. In terms of marker assisted selection (MAS), these potentially significantly correlated SNPs can be served as molecular markers for MAS. For example, for those yellow feathered broiler chickens that aim to cultivate longer keel length, larger body weight, and higher breast muscle percentage as breeding targets, early selection of chicken populations can be carried out based on the advantageous genotypes of keel length that we have screened, in order to accelerate the breeding process of keel length and improve breeding efficiency. By using MAS, it is possible to cultivate a new breed (matching line) of yellow feathered broilers with longer keel length, larger body weight, better chest muscle percentage that meets consumer needs in a relatively short period of time, thereby enhancing its market competitiveness.

Although this study identified multiple candidate genes and inferred their important roles in chicken keel development based on previous research, there is a lack of specific molecular mechanisms and direct functional verification of the regulation of keel growth and development by these genes. Next, we will investigate the impact of key candidate gene expression on the development of keel length, and determine the functions of key candidate genes through interference and overexpression experiments to construct a gene regulatory network. In addition, we will further validate the effectiveness of these SNP molecular markers targeting the dominant genotypes of keel length in this specialized strain and other populations, and apply them to keel length molecular marker assisted breeding.

Conclusions

We identified 10 SNPs potentially significantly associated with keel length through GWAS and other research methods, and analyzed their dominant genotypes. Through annotation of candidate genes for these SNP loci and previous related studies, 9 candidate genes were selected, among which CST3, ATP7B, OTOP1, COPS2 and IFT140 may represent key candidate genes affecting keel length development. This study provide scientific theoretical support for elucidating the molecular regulatory mechanisms of keel development and MAS of keel length.

Ethics statement and consent to participate

The animal experiments were performed in accordance with the protocol of the Animal Use Committee of the Chinese Ministry of Agriculture and were approved by the Animal Care and Use Committee at the Jiangsu Institute of Poultry Science (Approval ID: S20221023,Yangzhou, China). All experiments were performed in accordance with relevant guidelines and regulations.

CRediT authorship contribution statement

Y.J. Tu: Writing – review & editing, Writing – original draft, Data curation, Conceptualization. Y.F. Liu: Software, Resources. M. Zhang: Software, Data curation. X.J. Ju: Validation, Supervision. Y.J. Shan: Visualization, Methodology. G.G. Ji: Methodology, Formal analysis. J.T. Shu: Writing – review & editing, Funding acquisition.

Disclosures

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.