SUMMARY: Here, we present sbml2hyb, an easy-to-use standalone Python tool that facilitates the conversion of existing mechanistic models of biological systems in Systems Biology Markup Language (SBML) into hybrid semiparametric models that combine mechanistic functions with machine learning (ML). The so-formed hybrid models can be trained and stored back in databases in SBML format. The tool supports a user-friendly export interface with an internal format validator. Two case studies illustrate the use of the sbml2hyb tool. Additionally, we describe HMOD, a new model format designed to support and facilitate hybrid models building. It aggregates the mechanistic model information with the ML information and follows as close as possible the SBML rules. We expect the sbml2hyb tool and HMOD to greatly facilitate the widespread usage of hybrid modeling techniques for biological systems analysis. AVAILABILITY AND IMPLEMENTATION: The Python interface, source code and the example models used for the case studies are accessible at: https://github.com/r-costa/sbml2hyb. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

BACKGROUND: Non-negative linear combinations of elementary flux modes (EMs) describe all feasible reaction flux distributions for a given metabolic network under the quasi steady state assumption. However, only a small subset of EMs contribute to the physiological state of a given cell. RESULTS: In this paper, a method is proposed that identifies the subset of EMs that best explain the physiological state captured in reaction flux data, referred to as principal EMs (PEMs), given a pre-specified universe of EM candidates. The method avoids the evaluation of all possible combinations of EMs by using a branch and bound approach which is computationally very efficient. The performance of the method is assessed using simulated and experimental data of Pichia pastoris and experimental fluxome data of Saccharomyces cerevisiae. The proposed method is benchmarked against principal component analysis (PCA), commonly used to study the structure of metabolic flux data sets. CONCLUSIONS: The overall results show that the proposed method is computationally very effective in identifying the subset of PEMs within a large set of EM candidates (cases with ~100 and ~1000 EMs were studied). In contrast to the principal components in PCA, the identified PEMs have a biological meaning enabling identification of the key active pathways in a cell as well as the conditions under which the pathways are activated. This method clearly outperforms PCA in the interpretability of flux data providing additional insights into the underlying regulatory mechanisms.

Previous studies have reported the production of malformed virus-like-particles (VLP) in recombinant host systems. Here we computationally investigate the case of a large triple-layered rotavirus VLP (RLP). In vitro assembly, disassembly and reassembly data provides strong evidence of microscopic reversibility of RLP assembly. Light scattering experimental data also evidences a slow and reversible assembly untypical of kinetic traps, thus further strengthening the fidelity of a thermodynamically controlled assembly. In silico analysis further reveals that under favourable conditions particles distribution is dominated by structural subunits and completely built icosahedra, while other intermediates are present only at residual concentrations. Except for harshly unfavourable conditions, assembly yield is maximised when proteins are provided in the same VLP protein mass composition. The assembly yield decreases abruptly due to thermodynamic equilibrium when the VLP protein mass composition is not obeyed. The latter effect is more pronounced the higher the Gibbs free energy of subunit association is and the more complex the particle is. Overall this study shows that the correct formation of complex multi-layered VLPs is restricted to a narrow range of association energies and protein concentrations, thus the choice of the host system is critical for successful assembly. Likewise, the dynamic control of intracellular protein expression rates becomes very important to minimize wasted proteins.

BACKGROUND: The progress in the "-omic" sciences has allowed a deeper knowledge on many biological systems with industrial interest. This knowledge is still rarely used for advanced bioprocess monitoring and control at the bioreactor level. In this work, a bioprocess control method is presented, which is designed on the basis of the metabolic network of the organism under consideration. The bioprocess dynamics are formulated using hybrid rigorous/data driven systems and its inherent structure is defined by the metabolism elementary modes. RESULTS: The metabolic network of the system under study is decomposed into elementary modes (EMs), which are the simplest paths able to operate coherently in steady-state. A reduced reaction mechanism in the form of simplified reactions connecting substrates with end-products is obtained. A dynamical hybrid system integrating material balance equations, EMs reactions stoichiometry and kinetics was formulated. EMs kinetics were defined as the product of two terms: a mechanistic/empirical known term and an unknown term that must be identified from data, in a process optimisation perspective. This approach allows the quantification of fluxes carried by individual elementary modes which is of great help to identify dominant pathways as a function of environmental conditions. The methodology was employed to analyse experimental data of recombinant Baby Hamster Kidney (BHK-21A) cultures producing a recombinant fusion glycoprotein. The identified EMs kinetics demonstrated typical glucose and glutamine metabolic responses during cell growth and IgG1-IL2 synthesis. Finally, an online optimisation study was conducted in which the optimal feeding strategies of glucose and glutamine were calculated after re-estimation of model parameters at each sampling time. An improvement in the final product concentration was obtained as a result of this online optimisation. CONCLUSION: The main contribution of this work is a novel bioreactor optimal control method that uses detailed information concerning the metabolism of the underlying biological system. Moreover, the method allows the identification of structural modifications in metabolism over batch time.

The main aim of this study is to develop a bioprocess dynamic optimisation method that integrates bioreactor transport phenomena, with metabolic flux analysis (MFA). The central metabolic pathways of many biological systems with industrial interest are currently known. Bioreactor dynamic optimisation schemes could profit from the incorporation of this knowledge. A hybrid modelling methodology is presented that integrates the aforementioned concepts. The technique was successfully validated with data of a recombinant Baby Hamster Kidney (BHK-21) culture. The method allowed to identify the time evolution of intracellular metabolic fluxes and to relate this knowledge with bioreactor decision variables.