TY - JOUR TI - Linear-in-flux-expression (LIFE) approach to dynamic biological networks DO - https://doi.org/doi:10.7282/t3-k49a-yr31 PY - 2020 AB - This work analyzes the dynamics of three distinct classes of biological systems. The first is metabolic networks. The methodology named LIFE (Linear-in-Flux-Expression) was developed with the purpose of studying and analyzing large metabolic systems. With LIFE, the number of model parameters is reduced by accounting for correlations among the components of the system. These systems can be associated to graphs. General results on the stability of LIFE systems are discussed, particularly we formulate necessary conditions on the graph's structure to ensure the stability of the dynamics. Moreover, stability analysis from related fields, such as Markov chains, network flows, and compartmental systems, can also be applied. Control of LIFE systems through the addition of drugs as well as modifying intakes is discussed. A generalized graph object which incorporates hyperedges and uberedges is used to apply LIFE to metabolic networks, in particular to Mycobacterium tuberculosis (MTB). Results from LIFE simulations on MTB carbon metabolism are presented via simulations. Finally, the method allows us to rank 4-drug combinations in terms of their effectiveness in destabilizing MTB metabolic networks, thus killing the bacterium. The second class of systems is models for circadian rhythm. One of the essential characteristics of an authentic circadian clock is that the free-running period sustains an approximately24-hour cycle. The dynamics of the circadian clock is modified by an external stimulus, called a zeitgeber. This modification process is known as entrainment and operates to reset the phase and period of the circadian clock. When analyzing the phase of entrainment of many individuals, it is often assumed that an organism with a short period will have a phase advance, and a prolonged period will have a phase delay; however, this does not explain all known experimental data, so a Two-Step Entrainment model was developed. This work analyzes how parameters of the model affect the dynamics and presents results fitting the Two-Step Entrainment model to human data. The third class of systems consists of ecological networks. The interactions of species are often described via a network. Construction of networks in paleoecology is challenging due to the lack of observations of interactions, as well as biases in the preservation of species. The links of species in these networks must be inferred based on properties such as body size, similarities to living species, genetic information (when possible), and other known characteristics. Studying how paleo-networks have changed and adapted through time could assist in predicting how current ecological communities might react to environmental stressors. This work reconstructs networks from arthropod data found in rodent middens. The dynamics of these networks over 20,000 years is analyzed, and network metrics such as connectance are compared to modern networks. KW - Metabolic networks KW - Computational and Integrative Biology LA - English ER -