Animal samples and trait measurement
The DNA samples of 766 chickens were obtained from the following eight populations: Lushi chickens (LS, n = 39, 6 weeks), Ningdu chickens (ND, n = 95, 12 weeks), Wenchang chickens (WC, n = 65, 7 weeks), Qingyuan Partridge chickens (QY, n = 70, 7 weeks), Recessive White Rock chickens (RW, n = 55, 7 weeks), ISA Brown laying hen (ISA, n = 54, 20 weeks), Guangxi chickens (GX, n = 71, 12 weeks) and F2 population (F2, n = 360, 13 weeks). These DNA samples were all from the chicken breed resource library maintained in our laboratory. Among the eight different breeds, LS, ND, WC, QY and GX are domestic chicken breeds in China and RW and ISA are commercial broilers and layer hens, respectively. Additionally, the F2 resource population is a hybrid strain of RW and Xinghua (XH) chickens; XH chickens represent a slow-growing Chinese domestic chicken. In our laboratory, 2 mL of 5% pentobarbital was injected intraperitoneally into the chicken (No. 57–33-0; Chinese Academy of Sciences, Beijing Siyuan Technology Co., Ltd.). After 2–3 min, the chicken was sacrificed by bleeding through the carotid artery. Data records about economic traits, as well as detailed information on the measuring methods, were available for the F2 population, as previously described [30].
Twelve different tissues were obtained from four QY chickens. Additionally, the breast muscle of six embryonic periods (E10-15) and leg muscle of four embryonic periods (E12-15) were used to detect relative GNB1L expression.
cDNA synthesis and qRT–PCR
RNA was extracted using TRIzol (Takara, Dalian, China). Next, the RNA was reverse transcribed using the cDNA reverse transcription kit (Takara, Dalian, China) and then was subjected to PCR. Relative gene expression was calculated using the 2–ΔΔCt method, and significance was determined using ANOVA followed by Duncan's test. All the reactions were performed using three biological and technical repetitions. The relative expression of GNB1L using different tissues and embryo ages were analyzed by qRT–PCR. The qRT–PCR primers used for GNB1L and the internal control β-actin are listed in Table S3.
Indel detection and diversity analysis of different breeds
A 31-bp indel was identified in GNB1L from whole-genome resequencing data of ten XH and ten RW chickens (unpublished data). Genotyping of the GNB1L 31-bp indel was performed by PCR amplification and gel electrophoresis in eight diverse populations. Blood samples were used to extract DNA, and the final DNA concentration used for amplification was diluted to 50 ng/µL. The GNB1L PCR primers based on the genome are listed in Table S3. Each 15-µL PCR amplification volume contained 1 µL of DNA, 1.5 µL of primer, 7.5 µL 2×Taq Master mix (TSINGKE, Beijing, China), and 5 µL of double-distilled water. The PCR parameters were as follows: 95°C for 3 min, 35 cycles at 95°C for 30 s, 60°C for 30 s, 72°C for 30 s, and a final extension at 72°C for 10 min. The PCR products after amplification were separated by 3.0% gel electrophoresis.
The genotype and allele frequencies of the mutation were calculated directly in different breeds. Hardy–Weinberg equilibrium (HWE) was analyzed using the SHEsis website (http://analysis. biox.cn). Moreover, the allele numbers (Ne), genetic indices of heterozygosity (He), effective polymorphism information content (PIC) and population differentiation were analyzed using PopGene software (Version 1.3.1) [31, 32].
Transcription Factor Prediction
The transcription factors (TFs) in the 31-bp indel mutation of the 5′ UTR region of GNB1L were predicted using AliBaba software (Version 2.1) [24].
Statistics
Association analysis of the F2 population was performed using SPSS 22.0 software, and two different models were used in the analysis. All the growth traits used Model Ⅰ (Yijkl=µ+ Gi + Sj + Hk + fl + eijkl), and all the carcass traits used Model II (Yijkl=μ + Gi + Sj + Hk + fl + b (Wijkl -) + eijkl); the carcass weight served as a concomitant variable of Model II. Yijkl represents the observed value, µ is the overall population mean, fl is the fixed effect of family, Gi is the fixed effect of genotype, Hk is the fixed effect of hatch, Sj is the fixed effect of sex, b is the regression coefficient for carcass weight, is average slaughter weight, Wijkl represents the individual slaughter weight, and eijkl represents the random error in the two models. Significance was set at a P-value < 0.05, and Bonferroni’s test was performed for multiple comparisons [18].