of high fiber to low fiber in the treatments in a 72 day feeding trial. Metagenome sequencing was first carried out to derive the gene models for proteomics data search. MudPIT-based shot-gun proteomics and
enzyme assays were carried out to study biomass degradation mechanisms. The results highlighted that microbial diversity and enzyme profile change dramatically in response to the feeding content. Further
cluster analysis and protein co-regulatory network modeling of shot-gun proteomics data revealed that several groups of enzymes are coordinatively regulated in response to the high fiber biomass feeding. The
enzyme expression pattern correlates with the general enzyme activity in the rumen content. The co-regulatory network has highlighted the mechanisms for biomass degradation and the information has been
exploited to guide the reverse design of enzyme mixture and microbial strains. Based on the network modeling, we have cloned and characterized multiple enzymes, and developed the reverse design using
de novo enzyme mixture accordingly for efficient biomass utilization. Furthermore, we transformed several key cellullolytic and hemicellulytic enzymes to Saccharomyces cerevisiae to reduce enzyme load and
potentially enable a semi-CBP process for biofuel production. Overall, the study has revealed that system biology analysis of model natural biomass utilization systems like cattle rumen can serve as an
effective approach to guide the reverse design of enzyme mixtures, microbial strains, and potentially biorefinery processes.