Getting started¶
Enzyme Constraint (gecko) models work as regular cobra models with:
Proteins as additional metabolites.
Protein pseudorreactions, which simulate the production of the protein by the cell.
(Optional) Pool constraint: all proteins must be below the total quantified proteome.
The interface mirrors and depends on the cobrapy API.
from os.path import join, pardir
import geckopy
ROOT = pardir
DATA = join(ROOT, "tests", "data")
Import models¶
The models can be imported directly from SBML. Check the design chapter for more information on how proteins are identified.
ec_model = geckopy.io.read_sbml_ec_model(join(DATA, "eciML1515.xml.gz"))
Model derives from
cobra.core.model.
print(f"Model has {len(ec_model.reactions)} reactions and {len(ec_model.metabolites)} metabolites.")
Model has 4824 reactions and 2333 metabolites.
Models can also be written to the SBML file.
geckopy.io.write_sbml_ec_model(ec_model, join(DATA, "model_copy.xml"))
Python API: Protein isolation¶
Proteins are separated from metabolites and reactions in a different
container, proteins.
ec_model.slim_optimize()
0.876997313396373
ec_model.proteins.prot_P0A796
| Protein identifier | prot_P0A796 |
| Name | prot_P0A796 [cytosol] |
| Memory address | 0x0%x140634108785232 |
| Concentration | nan |
| Upper bound | 1000.0 |
| Mw | 0.0 |
| In 3 reaction(s) | PFKNo1 (62.00), PFK_2No1 (62.00), PFK_3No1 (62.00) |
Reactions are aware of their proteins.
ec_model.reactions.PFKNo1.proteins
{<Protein prot_P0A796 at 0x7fe7ee18f250>: -4.4803e-06}
Proteins are aware of their reactions.
ec_model.proteins.prot_P0A796.reactions
frozenset({<Reaction PFKNo1 at 0x7fe7ecc97b80>,
<Reaction PFK_2No1 at 0x7fe7ed25ce20>,
<Reaction PFK_3No1 at 0x7fe7ed413e50>})
Analogous to autoapi/cobra/index.html#cobra.reaction.flux.
ec_model.proteins.prot_P0A796.contribution
2.7670111370232322e-05
Fluxes are separated in the solution dataframe to avoid regexing:
# Fluxes are separated in the solution dataframe to avoid regexing:
solution_rxn, solution_prot = ec_model.optimize()
solution_rxn
| fluxes | reduced_costs | |
|---|---|---|
| EX_acgam_e | 0.000000 | -2.821258e-01 |
| EX_cellb_e | 0.000000 | -3.630018e-01 |
| EX_chol_e | 0.000000 | -2.633174e-02 |
| EX_pi_e | 0.000000 | 0.000000e+00 |
| EX_h_e | 8.058201 | 0.000000e+00 |
| ... | ... | ... |
| PUACGAMS_REVNo1 | 0.000000 | 0.000000e+00 |
| ARHGDx_REVNo1 | 0.000000 | 5.312591e-17 |
| UDPGPT_REVNo1 | 0.000000 | -5.551115e-17 |
| 4HTHRA_REVNo1 | 0.000587 | 0.000000e+00 |
| RHMND_REVNo1 | 0.000000 | 0.000000e+00 |
4824 rows × 2 columns
solution_prot
| fluxes | reduced_costs | |
|---|---|---|
| prot_P0A825 | 4.246198e-07 | 0.0 |
| prot_P75823 | 0.000000e+00 | -0.0 |
| prot_P0AEA8 | 0.000000e+00 | -0.0 |
| prot_P36553 | 3.195620e-06 | 0.0 |
| prot_P06715 | 1.642762e-07 | 0.0 |
| ... | ... | ... |
| prot_P77215 | 0.000000e+00 | -0.0 |
| prot_P0A8Y8 | 0.000000e+00 | -0.0 |
| prot_P76290 | 0.000000e+00 | -0.0 |
| prot_P16691 | 0.000000e+00 | -0.0 |
| prot_P32138 | 0.000000e+00 | -0.0 |
1259 rows × 2 columns
Kcats¶
The kcats can be inspected and manipulated from the
Protein object, as a regular dictionary.
These kcats are individual for every protein-reaction pair and correspond to the stoichiometric coefficient of the protein pseudometabolite in the reaction.
The units of the input are in \(\frac{1}{s}\).
This input is translated to \(h\) in the stoichiometric coefficient.
ec_model.proteins.prot_P0A796.kcats
{<Reaction PFKNo1 at 0x7fe7ecc97b80>: 61.99981648054322, <Reaction PFK_2No1 at 0x7fe7ed25ce20>: 61.99981648054322, <Reaction PFK_3No1 at 0x7fe7ed413e50>: 61.99981648054322}
ec_model.proteins.prot_P0A796.kcats["PFKNo1"]
61.99981648054322
ec_model.reactions.PFKNo1.metabolites[ec_model.proteins.prot_P0A796]
-4.4803e-06
ec_model.proteins.prot_P0A796.kcats["PFKNo1"] = 120
ec_model.reactions.PFKNo1.metabolites[ec_model.proteins.prot_P0A796]
-2.3148148148148148e-06
