Parameter Balancing - Short Introduction

What is parameter balancing?

The detailed description of the parameter balancing method can be obtained by the underlying publication (Lubitz et al. (2010)). Nevertheless, the essential knowledge for the practical usage can also be found on this page.

Parameter balancing helps you to find a consistent parameter set for a kinetic model, based on (i) kinetic constants and other data collected from experiments or the literature; (ii) constraints between biochemical quantities due to their definition or the laws of thermodynamics; (iii) assumptions about typical ranges, represented by prior distributions and pseudo values.

Our workflow starts from a given SBML model and a data table, balances the parameters, and inserts standard rate laws with the balanced parameters into the model (either replacing or completing the existing kinetic laws). As our standard rate laws, we use the modular rate laws (see Liebermeister et al. (2010)).

The parameter balancing workflow - step by step

The parameter balancing workflow proceeds as follows:

  1. Choose an SBML model from the model list and click the "Parameter balancing" symbol.

  2. Upload a data table. For the file format, please see the instructions for file formats. If no data table is uploaded, parameter balancing will be based on priors and pseudo values only.

  3. Preprocess data. This screen shows the data table you uploaded before. You can choose options for filtering data (for source organism), completion of missing standard errors, and unit conversion (currently, only the conversion from molecules/cell to mM is supported).

  4. Prepare parameter balancing. On this screen, you see the quantities needed for your model. Missing quantities are shown as blank lines and duplicate available values have been averaged (for non-energy quantities, on log scale) using the formulas σ2 = 1/(Σ (1/σ2j)) and x = σ2 Σ (xj2j). For the next step, you can edit here the means and widths of prior distributions and pseudo values; choose between actual parameter balancing or different ways to directly insert prior and pseudo values into the data table.

  5. Insert rate laws with balanced parameters into SBML model; choose type of rate law, options for insertion; optionally, choose number of data sets to be sampled from the posterior. The resulting model is added automatically to your model list.

  6. Optionally: download the sampled models and data sets (only if parameter sets have been sampled before).

Demo examples

Here you can find a number of example files (SBML models and data tables for parameter balancing).

How does parameter balancing work internally?

Parameter balancing employs Bayesian estimation to determine a consistent set of all model parameters. To use it efficiently, it is good to know about some of its details. For technical reasons, all quantities are internally converted to natural scaling. This means that for energy quantities (in kJ/mol), we keep the original values while for all other quantities, we take the natural logarithms. Furthermore, we distinguish between basic quantities and derived quantities (which are uniquely determined by the basic quantities). See the overview of all quantities considered.
During balancing, we integrate information from data (values and standard errors), prior distributions (typical values and spread for basic quantities) and pseudo values (typical values and spread for derived quantities). All these values and spreads are represented by normal distributions (priors, data with standard errors, pseudo values, and posteriors) for the naturally scaled quantities. When converting back to non-logarithmic values, we obtain log-normal distributions, which makes it crucial to distinguish between median and mean values. Eventually, the median values (which are more realistic and guaranteed to satisfy the relevant constraints) are inserted into the model.

Links

  1. Back to semanticSBML 2.0 Main
  2. Instructions for file formats
  3. Example files
  4. Overview of quantities considered