Limits...
Partitioning of genomic variance reveals biological pathways associated with udder health and milk production traits in dairy cattle.

Edwards SM, Thomsen B, Madsen P, Sørensen P - Genet. Sel. Evol. (2015)

Bottom Line: Several biological pathways that were significantly associated with health and production traits were identified in dairy cattle; i.e. the markers linked to these pathways explained more of the genomic variance and provided a better model fit than 95 % of the randomly sampled gene groups.Our results show that immune related pathways are associated with production traits, and that pathways that include a causal marker for production traits are identified with our procedure.We are confident that the LMM approach provides a general framework to exploit and integrate prior biological information and could potentially lead to improved understanding of the genetic architecture of complex traits and diseases.

View Article: PubMed Central - PubMed

Affiliation: Center for Quantitative Genetics and Genomics, Department of Molecular Biology and Genetics, Aarhus University, Blichers Allé 20, P.O. Box 50, Tjele, DK-8830, Denmark. Stefan.Hoj-Edwards@roslin.ed.ac.uk.

ABSTRACT

Background: We have used a linear mixed model (LMM) approach to examine the joint contribution of genetic markers associated with a biological pathway. However, with these markers being scattered throughout the genome, we are faced with the challenge of modelling the contribution from several, sometimes even all, chromosomes at once. Due to linkage disequilibrium (LD), all markers may be assumed to account for some genomic variance; but the question is whether random sets of markers account for the same genomic variance as markers associated with a biological pathway?

Results: We applied the LMM approach to identify biological pathways associated with udder health and milk production traits in dairy cattle. A random gene sampling procedure was applied to assess the biological pathways in a dataset that has an inherently complex genetic correlation pattern due to the population structure of dairy cattle, and to linkage disequilibrium within the bovine genome and within the genes associated to the biological pathway.

Conclusions: Several biological pathways that were significantly associated with health and production traits were identified in dairy cattle; i.e. the markers linked to these pathways explained more of the genomic variance and provided a better model fit than 95 % of the randomly sampled gene groups. Our results show that immune related pathways are associated with production traits, and that pathways that include a causal marker for production traits are identified with our procedure. We are confident that the LMM approach provides a general framework to exploit and integrate prior biological information and could potentially lead to improved understanding of the genetic architecture of complex traits and diseases.

No MeSH data available.


Related in: MedlinePlus

QQ-plot of the observed likelihood ratios (LR) of random gene groups vs. theoretical -distribution showing that they are skewed towards higher values than . LR displayed for the traits Mastitis 1.1 (top) and Fat yield (bottom), conditional on whether the gene groups contain one of the DGAT1 genes (left/right) and group size (colour). Please note that the range of the y-axes differs between the top and bottom
© Copyright Policy - open-access
Related In: Results  -  Collection

License 1 - License 2
getmorefigures.php?uid=PMC4499908&req=5

Fig2: QQ-plot of the observed likelihood ratios (LR) of random gene groups vs. theoretical -distribution showing that they are skewed towards higher values than . LR displayed for the traits Mastitis 1.1 (top) and Fat yield (bottom), conditional on whether the gene groups contain one of the DGAT1 genes (left/right) and group size (colour). Please note that the range of the y-axes differs between the top and bottom

Mentions: Distributions of LR statistics are compared in Fig. 2 where the random gene groups are presented according to group size in intervals of 10000 markers. This revealed that the distributions for group sizes smaller than 10 000 markers differed in a trait-dependent way. For fat yield, the LR distribution was skewed towards larger values, whereas the LR distribution for health traits was slightly skewed towards smaller values, when the gene group included one of the DGAT1 genes. We also found a weak, but significant increase in the 95th percentile of the LR statistics for different group sizes. Therefore, to better account for the influence of group size in the LRT, we used the quantile regression approach [34–36] to determine a 95 % cut-off adjusted for group size and the DGAT1 genes in the gene groups.Fig. 2


Partitioning of genomic variance reveals biological pathways associated with udder health and milk production traits in dairy cattle.

Edwards SM, Thomsen B, Madsen P, Sørensen P - Genet. Sel. Evol. (2015)

QQ-plot of the observed likelihood ratios (LR) of random gene groups vs. theoretical -distribution showing that they are skewed towards higher values than . LR displayed for the traits Mastitis 1.1 (top) and Fat yield (bottom), conditional on whether the gene groups contain one of the DGAT1 genes (left/right) and group size (colour). Please note that the range of the y-axes differs between the top and bottom
© Copyright Policy - open-access
Related In: Results  -  Collection

License 1 - License 2
Show All Figures
getmorefigures.php?uid=PMC4499908&req=5

Fig2: QQ-plot of the observed likelihood ratios (LR) of random gene groups vs. theoretical -distribution showing that they are skewed towards higher values than . LR displayed for the traits Mastitis 1.1 (top) and Fat yield (bottom), conditional on whether the gene groups contain one of the DGAT1 genes (left/right) and group size (colour). Please note that the range of the y-axes differs between the top and bottom
Mentions: Distributions of LR statistics are compared in Fig. 2 where the random gene groups are presented according to group size in intervals of 10000 markers. This revealed that the distributions for group sizes smaller than 10 000 markers differed in a trait-dependent way. For fat yield, the LR distribution was skewed towards larger values, whereas the LR distribution for health traits was slightly skewed towards smaller values, when the gene group included one of the DGAT1 genes. We also found a weak, but significant increase in the 95th percentile of the LR statistics for different group sizes. Therefore, to better account for the influence of group size in the LRT, we used the quantile regression approach [34–36] to determine a 95 % cut-off adjusted for group size and the DGAT1 genes in the gene groups.Fig. 2

Bottom Line: Several biological pathways that were significantly associated with health and production traits were identified in dairy cattle; i.e. the markers linked to these pathways explained more of the genomic variance and provided a better model fit than 95 % of the randomly sampled gene groups.Our results show that immune related pathways are associated with production traits, and that pathways that include a causal marker for production traits are identified with our procedure.We are confident that the LMM approach provides a general framework to exploit and integrate prior biological information and could potentially lead to improved understanding of the genetic architecture of complex traits and diseases.

View Article: PubMed Central - PubMed

Affiliation: Center for Quantitative Genetics and Genomics, Department of Molecular Biology and Genetics, Aarhus University, Blichers Allé 20, P.O. Box 50, Tjele, DK-8830, Denmark. Stefan.Hoj-Edwards@roslin.ed.ac.uk.

ABSTRACT

Background: We have used a linear mixed model (LMM) approach to examine the joint contribution of genetic markers associated with a biological pathway. However, with these markers being scattered throughout the genome, we are faced with the challenge of modelling the contribution from several, sometimes even all, chromosomes at once. Due to linkage disequilibrium (LD), all markers may be assumed to account for some genomic variance; but the question is whether random sets of markers account for the same genomic variance as markers associated with a biological pathway?

Results: We applied the LMM approach to identify biological pathways associated with udder health and milk production traits in dairy cattle. A random gene sampling procedure was applied to assess the biological pathways in a dataset that has an inherently complex genetic correlation pattern due to the population structure of dairy cattle, and to linkage disequilibrium within the bovine genome and within the genes associated to the biological pathway.

Conclusions: Several biological pathways that were significantly associated with health and production traits were identified in dairy cattle; i.e. the markers linked to these pathways explained more of the genomic variance and provided a better model fit than 95 % of the randomly sampled gene groups. Our results show that immune related pathways are associated with production traits, and that pathways that include a causal marker for production traits are identified with our procedure. We are confident that the LMM approach provides a general framework to exploit and integrate prior biological information and could potentially lead to improved understanding of the genetic architecture of complex traits and diseases.

No MeSH data available.


Related in: MedlinePlus