Cattle are the main source of methane emissions from livestock, which have a 34-times higher greenhouse effect than CO2. In the Global Methane Pledge, 105 countries committed to reducing methane emissions by 30% from 2020 levels during the following 10 years. A lack of understanding of the microbiological processes producing rumen methane emissions makes existing mitigation strategies inadequate to meet this challenge. Multi-omics technologies are overcoming the limits of culture-based microbial research and revealing microbial mechanisms behind the host phenotype. Traditional bioinformatic methodologies fail with microbiome omics-data due to limited computing efficiency and unpredictable integration. Multidimensionality and multi-omics data lack appropriate computational methods. This Thesis proposed integrative research frameworks for three topics. 1) Metabolite identification. Knowledge-driven network analysis based on enzymatic reactions, non-linear and linear correlations among microbial genes and NMR spectroscopy. This study was the first to infer the corresponding metabolites based on microbial genetic information. The accuracy of metabolite quantification was validated by comparison between methods and recommendations for relevant data pre-processing/pre-treatment were made. The goal of completing the first metabolite dataset with 115 metabolites quantified and identified was achieved. 2) Biomarker identification. This Thesis developed multilayer networks based on quantitative and knowledge association of multi-omics data by the combination of network approach and statistical ranking inference. The models captured existing biomarkers in key modules and identify new biomarkers for specific biological processes. The success of the integrated frameworks was demonstrated by consistent results at the three levels of microbial genes, metabolites, and microbial communities associated with methane function. 3) Multi-omics systematic analysis. The Integrated Multi-dimensional Omics Framework (IMOF) for host-microbiome data was proposed to illustrate the dynamic development of rumen methane emissions by the microbiome in response to a variety of environmental conditions. Our studies have made significant contribution to knowledge about microbial mechanisms involved in the between rumen methane emission and rumen environment.
- artificial intelligence algorithm
- methane emission
- rumen microbiome
- multi-omics
High-dimensional multi-omics data integrative analysis of the rumen microbiome associated with bovine methane emissions
Wang, M. (Author). Apr 2023
Student thesis: Doctoral Thesis