QTCCm: From Quantitative Trait Correlation to Causation in dairy cattle – using transcriptome and metabolome data

• Applicants: Professor Dr. Jörn Bennewitz; Professor Dr. Jens Tetens; Professor Dr. Georg Thaller
• Subject Area: Animal Breeding, Animal Nutrition, Animal Husbandry
• Term: since 2020
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 448536632

The breeding goal in dairy cattle breeding includes performance traits (e.g. milk yield) and fertility and health traits, and longevity. Modern genomic tools applied to large data sets and augmented with external information about genome sequence variants allows for the comprehensive dissection of the genetic architecture of quantitative traits and, beyond that, of the shared architecture of multiple traits. In the first funding period we were able to map the genetic correlation between these trait complexes down to chromosomal regions and functional annotated genome variant subsets. Further, it was possible to detect causal structures underlying these correlations. Starting from these very convincing results, the overall aim of the second funding period is to gain deep insight into the interrelationships between milk production, feed efficiency, and health, fertility and longevity traits (the latter three are summarised as HFL-traits in the following) in German Holsteins dairy cattle. The following hypotheses will be tested. (i) The assessment of the genetic correlation between milk and HFL traits using longitudinal sampled transcriptome, metabolome, and trait data quantifies the causative contribution of known pathways and detects novel metabolic pathways linking these traits complexes. (ii) The detected synergistic and antagonistic genetic correlation pattern between milk and HFL traits across the genome and functional genomic regions can be fine-mapped down to functional genes together with regulatory elements, and CV using transcriptomic and metabolic data. (iii) CV for genetic correlation between milk production and HFL traits can be classified in horizontal- and vertical-pleiotropic variants using causal inference methods. These hypotheses will be tested by establishing an animal experiment on the research farm Karkendamm, where 250 milking cows will be longitudinal sampled within a lactation for various records including blood transcriptome and metabolome probes. These data will be connected with the existing and very powerful QTCC data set with 200k cows with imputed sequence data and a large panel of quantitative traits. The data will be analysed using latest genomic and statistical methods, paying special attention on holistic approaches using mixed linear models and their derivates. The results are important to understand the nature of the interrelationship. This is needed to understand the consequences of selection decisions on correlated traits and to develop multi-trait genomic selection schemes that avoid undesirable side effects, to enhance the understanding of so-called production diseases and fertility problems, and to assess the effects of putative external interventions on a causative trait with the aim to avoid deleterious causative effects compromising animal health and welfare.

• DFG Programme: Research Grants