Efficient parameterization of large-scale dynamic models based on relative measurements

Abstract

Motivation Mechanistic models of biochemical reaction networks facilitate the quantitative understanding of biological processes and the integration of heterogeneous datasets. However, some biological processes require the consideration of comprehensive reaction networks and therefore large-scale models. Parameter estimation for such models poses great challenges, in particular when the data are on a relative scale. Results Here, we propose a novel hierarchical approach combining (i) the efficient analytic evaluation of optimal scaling, offset and error model parameters with (ii) the scalable evaluation of objective function gradients using adjoint sensitivity analysis. We evaluate the properties of the methods by parameterizing a pan-cancer ordinary differential equation model (>1000 state variables, >4000 parameters) using relative protein, phosphoprotein and viability measurements. The hierarchical formulation improves optimizer performance considerably. Furthermore, we show that this approach allows estimating error model parameters with negligible computational overhead when no experimental estimates are available, providing an unbiased way to weight heterogeneous data. Overall, our hierarchical formulation is applicable to a wide range of models, and allows for the efficient parameterization of large-scale models based on heterogeneous relative measurements. Availability and implementation Supplementary code and data are available online at http://doi.org/10.5281/zenodo.3254429 and http://doi.org/10.5281/zenodo.3254441. Supplementary information Supplementary data are available at Bioinformatics online.

Parameterization of mechanistic models from qualitative data using an efficient optimal scaling approach

“…As the surrogate data are the solution to the optimization problem (6), their sensitivity is the sensitivity of this optimal solution with respect to the parameters θ . The availability of gradient information would allow for efficient hierarchical optimization algorithms similar to results for relative quantitative data (Weber et al 2011;Loos et al 2018;Schmiester et al 2019).…”

PEtab—Interoperable specification of parameter estimation problems in systems biology

“…However, some methods tailored to specific problems require additional information to estimate the unknown parameters. To acknowledge this, we allow for optional application-specific extensions in addition to the required columns in the PEtab files, e.g., if some parameters can be calculated analytically using hierarchical optimization approaches [18].…”

A data-driven computational model enables integrative and mechanistic characterization of dynamic macrophage polarization

“…Although in our analyses we selected an in vitro experimental condition of PAD to demonstrate model utility and its potential translational application, the mechanistic setup of our model in terms of the pathways and mechanisms included is equally important for macrophages under general as well as other disease-specific scenarios. Still, the model certainly does not account for the full biology (an innate limitation even for large-scale models), as it is practically infeasible, due to relatively limited experimental data and high computational cost, to detail and calibrate every known pathway and mechanism that regulate macrophage polarization within a single modeling study while maintaining a high degree of model performance accuracy (Frohlich et al, 2018;Bouhaddou et al, 2018;Schmiester et al, 2020). Thus, our current model and its formulation can instead serve as a high-quality mechanistic computational platform that can be accordingly and continuously expanded and enriched with additional pathway details to further investigate macrophage functions in specific disease areas of interest, e.g., TLR pathways in various infectious disease settings (O'neill et al, 2009), CD47/SIRPa axis in macrophage-mediated cancer immunotherapy (Weiskopf, 2017), and cellular metabolic pathways in nonalcoholic fatty liver disease (Oates et al, 2019).…”