Enzyme-constrained models (ec-Human2)

Work in progress

This page describes the enzyme-constrained version of Human-GEM (ec-Human2), which is being finalized and will be released soon. The instructions below are a draft and are likely to change.

Enzyme-constrained GEMs (ecGEMs) augment a metabolic model with enzyme kinetics (k_cat values) and a limited enzyme (protein) pool, so that reaction fluxes are additionally constrained by the abundance and turnover of the enzymes that catalyze them. This typically shrinks the feasible flux space and improves quantitative predictions. For Human-GEM, an enzyme-constrained version (ec-Human2) is generated with the GECKO method.

As described in Luo et al. (2026) PNAS, ec-Human2 improved flux predictions (e.g., tighter flux variability in NCI-60 cell lines) and the reproduction of inborn errors of metabolism, and it forms the basis of the enzyme-constrained whole-body simulations (see Whole-body models).

MATLAB (GECKO)

ec-Human2 is built and analysed with the GECKO Toolbox (v3). See the GECKO documentation for the full workflow of constructing an ecModel from a reference GEM such as Human-GEM.

Python (geckopy)

Experimental

geckopy is a Python implementation of GECKO and is under active development; it is not yet on PyPI. Install it from GitHub main:

pip install "git+https://github.com/SysBioChalmers/geckopy@main"

A full worked example (building ec-Human2, enzyme-constrained FBA/FVA, and inspecting enzyme usage) will be added here once ec-Human2 is released.