Figure 1.
The animals in blue boxes were phenotyped and genotyped and included in the analyses. G: Generation; P: Parity. Please refer to references [2], [10] for the population structure, which defines generation and parity.
Figure 2.
Clustering of the residual feed intake (RFI) population by the PLINK software.
Figure A indicates identical-by-state (IBS) clustering. The X and Y axes represent IBS mean and IBS variance. Figure B shows multi-dimensional clustering. The X and Y-axes indicate dimensions 1 and 2, respectively. The black color cluster includes animals from generations 3–8 of the Low RFI line, the green cluster includes animals from generations 4–8 of the High RFI line, and the red cluster includes animals from generations 1 and 2 of the Low RFI line and generation 0, which is the founder population from which RFI selection lines originated.
Figure 3.
Allele frequency differences between the Low and High residual feed intake lines at generation 8.
The X axis indicates in different colors from left to right, SNP locations from chromosomes 1 to X, unassigned contigs, Y, and completely unmapped SNPs, using Sus scrofa genome build10.2. The Y axis represents the minus log of the P-value for the allele frequency difference between the two lines for each SNP,. The dashed line shows the P-value threshold. SSC: Pig chromosome; FDR: False discovery rate; KCNJ15: Potassium inwardly rectifying channel, subfamily J, member 15; ELOVL2: Elongation of very long chain fatty acids 2; TFAP2A: Transcription factor AP-2 (Activating enhancer binding protein 2)—alpha; GPX2: Glutathione peroxidase 2.
Figure 4.
Allele frequency differences between the residual feed intake selection lines for the significant SNPs in each generation.
Parts A, B and C show the allele frequency differences for the SNPs near to the KCNJ15 gene on SSC13 (ASGA0060074), near the ELOVL2 and TFAP2A gene on SSC7 (ASGA0030976), and near the GPX2 gene on SSC7 (ALGA0043495), respectively. KCNJ15: potassium inwardly rectifying channel, subfamily J, member 15; ELOVL2: elongation of very long chain fatty acids 2; TFAP2A: transcription factor AP-2 (activating enhancer binding protein 2)—alpha; GPX2: glutathione peroxidase 2. The X axes represent generations and Y-axes show allele frequencies.
Table 1.
SNPs with significant (P value<0.0001 and FDR P value<0.05) allele frequency differences in generation 8 between Low and High RFI selection lines.
Figure 5.
Whole genome association studies for residual feed intake (RFI).
Part A depicts association analyses performed by the PLINK software for each SNP. The X axis shows SNPs across chromosomes 1 to X, unassigned contigs, Y and completely unmapped SNP. The Y axis represents the negative logarithm of the P-values corrected for genomic control. Each spot is a SNP. The green color SNPs are those located in 1 Mb window regions that explain more than 0.2% of genetic variance in part B. Part B illustrates results from the Bayes B model averaging approach used in the GenSel software. Different colors on the X axis indicate genome wide 1 Mb SNP windows from chromosome 1 to X, unassigned contigs, Y and completely unmapped SNP. The markers from completely unmapped and unassigned contigs were not included in the cumulative genetic variance. The Y axis represents percent genetic variance explained by each 1 Mb window. Part C shows association analyses with the PLINK software based on haplotypes, which were derived for 1 Mb windows that explained a higher than 0.2% of genetic variance in part B. The X axis shows chromosomal positions of the haplotypes. The Y axis shows the negative logarithm of the P-values corrected by genomic control. The arrows in parts A, B and C show the similarities in the significant locations. The 1 Mb windows that explained a higher than 0.2% percent of genetic variance in the Bayesian analyses and/or were significant in the PLINK analyses were considered to be important putative QTL for RFI. GNG4: guanine nucleotide binding protein 4; GLP1R: glucagon-like peptide 1 receptor; CDKAL1: cyclin-dependent kinase 5 regulatory subunit associated protein 1-like 1.
Table 2.
Important candidate QTL regions associated with the residual feed intake (RFI) by 1 Mb SNP windows.
Figure 6.
Whole genome association studies for average daily feed intake (ADFI).
Part A depicts association analyses performed by the PLINK software for each SNPs. The X axis shows SNPs across chromosomes 1 to X, unassigned contigs, Y and completely unmapped SNP. The Y axis represents the negative logarithm of the P values corrected for genomic control. Each spot is a SNP. The green color SNPs are those located in 1 Mb window regions that explain more than 0.2% of genetic variance in part B. Part B illustrates results from the Bayes B model averaging approach used in the GenSel software. Different colors on the X axis indicate genome wide 1 Mb SNP windows from chromosome 1 to X, unassigned contigs, Y and completely unmapped SNP. The markers from completely unmapped and unassigned contigs were not included in the cumulative genetic variance. The Y axis represents percent genetic variance explained by each 1 Mb window. Part C shows association analyses with the PLINK software based on haplotypes, which were derived for 1 Mb windows that explained a higher than 0.2% of genetic variance in part B. The X axis depicts chromosomal positions of haplotypes. The Y axis shows the negative logarithm of the P values corrected by genomic control. The arrows in parts A, B and C show the similarities in significant locations of the associated SNPs, SNP windows and their haplotypes. The 1 Mb windows that explained higher than 0.2% percent genetic variance in GenSel analyses and/or were significant in the PLINK analyses were considered to be important putative QTL for ADFI. SGMS1: spingomyelin synthase 1; CBLN4: cerebellin 4; KCNK1: potassium channel, subfamily K, member 1; MC4R: melanocortin 4 receptor; PAQR5: progestin and adipoQ receptor family member V; GNG4: guanine nucleotide binding protein 4; SORCS3: sortilin-related vps10 domain containing receptor 3.
Figure 7.
Whole genome association analyses for average daily gain (ADG).
Part A depicts association analyses performed by the PLINK software for each SNP. The X axis shows SNPs across chromosomes SSC1 to X, unassigned contigs, Y and completely unmapped SNP. The Y axis contains the negative logarithm of the P values adjusted for genomic control. Each spot is a SNP. The green color SNPs are those located in 1 Mb window regions that explain more than 0.2% of genetic variance in part B. Part B illustrates results from the Bayes B model averaging approach used in the Gensel software. Different colors on the X axis indicate genome wide 1MB SNP windows from chromosome 1 to X, unassigned contigs and completely unmapped SNP. The markers from completely unmapped and unassigned contigs were not included in the cumulative genetic variance. The Y axis represents percent genetic variance explained by each 1 Mb window. Part C shows association analyses with the PLINK software based on haplotypes, which were derived for 1 Mb windows that explained a higher than 0.2% of genetic variance in part B. The X axis has haplotypes on specific chromosomes. The Y axis shows the negative logarithm of the P values corrected by genomic control. The arrows in parts A, B and C show the similarities in significant locations of the associated SNPs, SNP windows and their haplotypes. The 1 Mb windows that explained higher than 0.2% percent genetic variance in Gensel analyses and/or were significant in the PLINK analyses were considered to be important putative QTL for ADG. MC4R: melanocortin 4 receptor; TGFBI: transforming growth factor beta induced protein ig-h3; PGM1: phosphoglucomutase 1; GPR81: G-protein coupled receptor 81; U6: spliceosomal RNA.
Figure 8.
Whole genome association studies for back fat (BF).
Part A depicts association analyses performed by the PLINK software for each SNP. The X axis shows SNPs across chromosomes SSC1 to X, unassigned cintigs, Y and completely unmapped SNP. The Y axis contains the negative logarithm of the P values adjusted for genomic control. Each spot is a SNP. The green color SNPs are those located in 1 Mb window regions that explain more than 0.2% of genetic variance in part B. Part B illustrates results from the Bayes B model averaging approach used in the Gensel software. Different colors on the X axis indicate genome wide 1 Mb SNP windows from chromosome 1 to X, unassigned contigs and completely unmapped SNP. The markers from completely unmapped and unassigned contigs were not included in the cumulative genetic variance. The Y axis represents percent genetic variance explained by each 1 Mb window. Part C shows association analyses with the PLINK software based on haplotypes, which were derived for 1 Mb windows that explained a higher than 0.2% of genetic variance in part B. The X axis has haplotypes on specific chromosomes. The Y axis shows the negative logarithm of the P values corrected by genomic control. The arrows in parts A, B and C show the similarities in significant locations of the associated SNPs, SNP windows and their haplotypes. The 1 Mb windows that explained higher than 0.2% percent genetic variance in Gensel analyses and/or were significant in the PLINK analyses were considered to be important putative QTL for BF. ACOXL: acyl-coenzyme A oxidase-like; AEBP1: adipocyte enhancer binding protein 1; Env: envelope protein.
Figure 9.
Whole genome association studies for loin muscle area (LMA).
Part A depicts association analyses performed by the PLINK software for each SNP. The X axis shows SNPs across chromosomes SSC1 to X, unassigned contigs, Y and unmapped SNP. The Y axis contains the negative logarithm of the P values adjusted for genomic control. Each spot is a SNP. The green color SNPs are those located in 1 Mb window regions that explain more than 0.2% of genetic variance in part B. Part B illustrates results from the Bayes B model averaging approach used in the Gensel software. Different colors on the X axis indicate genome wide 1 Mb SNP windows from chromosome 1 to X, unassigned contigs, Y and completely unmapped SNP. The markers from completely unmapped and unassigned contigs were not included in the cumulative genetic variance. The Y axis represents percent genetic variance explained by each 1 Mb window. Part C shows association analyses with the PLINK software based on haplotypes, which were selected from the 1 Mb windows that explained a higher than 0.2% of genetic variance in part B. The X axis has haplotypes on specific chromosomes. The Y axis shows the negative logarithm of the P values corrected by genomic control. The arrows in parts A, B and C show the similarities in significant locations of the associated SNPs, SNP windows and their haplotypes. The 1 Mb windows that explained higher than 0.2% percent genetic variance in Gensel analyses and/or were very highly significant even after genomic control followed by FDR (P value<0.05) in the PLINK analyses were considered to be important QTL for LMA. KLHL31: kelch like 31.