请问什么是reference panel？它在snp基因分型中有什么作用

请问什么是reference panel？它在snp基因分型中有什么作用

请根据下面的研究内容，列出详细的技术路线，如下：The phenotype and genotype data of 21 traits were collected from 2,622 individuals over the past decade, including our newly collected data from the above 120 cattle. The summary of phenotype records is shown in Table S1. All 2,622 individuals were genotyped using the Illumina Bovine 770K Bead chip. The SNPs with minor allele frequencies < 0.05, genotype call rates < 90%, located in non-autosome and significant Hardy–Weinberg disequilibrium at 1 × 10−6
, as well as samples with call rates < 90% were removed from the downstream analysis using PLINK 1.90 [33]. After quality control, a total of 590,065 variants remained. We then imputed the SNPs to sequence variants level based on a multiple breeds reference panel by Beagle 5.4 [34]. The reference panel consists 1,847 individuals (including 113 Simmental cattle) and was downloaded from https://www.ebi.ac.uk/ena/browser/view/ERZ1738264 [35]. We removed variants with MAF < 0.05 or dosage R-squared (DR2) < 0.8. After quality control (same as above), 8,221,244 autosomal variants were obtained for GWAS and eQTL mapping. The average DR2 of the imputed variants was 0.92. Before performing GWAS, the phenotype was adjusted by year, sex, age, and the first two principal components (PCs) of the genotype generated by PLINK and normalized by rank-transformation using the transform function in GenABEL [36]. We performed an association test for each SNP based on a linear mixed model:

yy=XXββ+uu+εε,
where y is adjusted phenotype; X is a vector of genotypes of a variant at the locus tested; β is the effect size of the variant; u is a vector of random polygenic effects ~ N (0, GGσσ2gg
), where G is genomic relationship matrix constructed from all variants; ε is a vector of residual errors. Variance component estimation via restricted maximum likelihood (REML) analysis was implemented in GCTA software [37]. We used P-value < 5 × 10−8 as significance thresholds of GWAS for all traits. We compared GWAS results with Cattle QTLdb (release 47, Apr 25, 2022) [1]. The QTLs that were within ± 100 kb of a QTL/association for the same trait(s) of the Cattle QTLdb were treated as the newly detected QTLs.

The phenotype and genotype data of 21 traits were collected from 2,622 individuals over the past decade, including our newly collected data from the above 120 cattle. The summary of phenotype records is shown in Table S1. All 2,622 individuals were genotyped using the Illumina Bovine 770K Bead chip. The SNPs with minor allele frequencies < 0.05, genotype call rates < 90%, located in non-autosome and significant Hardy–Weinberg disequilibrium at 1 × 10−6
, as well as samples with call rates < 90% were removed from the downstream analysis using PLINK 1.90 [33]. After quality control, a total of 590,065 variants remained. We then imputed the SNPs to sequence variants level based on a multiple breeds reference panel by Beagle 5.4 [34]. The reference panel consists 1,847 individuals (including 113 Simmental cattle) and was downloaded from https://www.ebi.ac.uk/ena/browser/view/ERZ1738264 [35]. We removed variants with MAF < 0.05 or dosage R-squared (DR2) < 0.8. After quality control (same as above), 8,221,244 autosomal variants were obtained for GWAS and eQTL mapping. The average DR2 of the imputed variants was 0.92. Before performing GWAS, the phenotype was adjusted by year, sex, age, and the first two principal components (PCs) of the genotype generated by PLINK and normalized by rank-transformation using the transform function in GenABEL [36]. We performed an association test for each SNP based on a linear mixed model:

yy=XXββ+uu+εε,
where y is adjusted phenotype; X is a vector of genotypes of a variant at the locus tested; β is the effect size of the variant; u is a vector of random polygenic effects ~ N (0, GGσσ2gg
), where G is genomic relationship matrix constructed from all variants; ε is a vector of residual errors. Variance component estimation via restricted maximum likelihood (REML) analysis was implemented in GCTA software [37]. We used P-value < 5 × 10−8 as significance thresholds of GWAS for all traits. We compared GWAS results with Cattle QTLdb (release 47, Apr 25, 2022) [1]. The QTLs that were within ± 100 kb of a QTL/association for the same trait(s) of the Cattle QTLdb were treated as the newly detected QTLs.

请问什么是SNP插补？它有什么作用

请问什么是average DR2？它有什么作用

下面是一段研究方法，如下：Before performing GWAS, the phenotype was adjusted by year, sex, age, and the first two principal components (PCs) of the genotype generated by PLINK and normalized by rank-transformation using the transform function in GenABEL. 请问其中的基因型的前两个主成分指的是什么？

GWAS中调整表型后，使用GenABEL中的transform函数进行秩转换，请问这是为什么？有什么作用？

GWAS中调整表型后，使用GenABEL中的transform函数进行秩转换，请问这是为什么？有什么作用？