search

SysBioChalmers / yeast-GEM

The consensus GEM for Saccharomyces cerevisiae
http://sysbiochalmers.github.io/yeast-GEM/
Creative Commons Attribution 4.0 International
96 stars 46 forks source link

feat: add metDeltaG and rxnDeltaG fields #330

Closed cheng-yu-zhang closed 1 year ago

cheng-yu-zhang commented 1 year ago

Main improvements in this PR:

Add physiological status ΔG for rxns and mets from Modelseed and yetfl. All the related data and code is in ../yeast-GEM/deltaG.

  1. ../yeast-GEM/deltaG/yeast-GEM_with_G/yeast.mat is the final model with ΔG. ΔG is store in fieldyeast.metsdeltaG and yeast.rxnsdeltaG.
  2. ../yeast-GEM/deltaG/model_from_yetfl/yeast8_thermo_curated.mat is the model from yetfl containing ΔG data.https://github.com/EPFL-LCSB/yetfl
  3. ../yeast-GEM/deltaG/met_G.csv" contains mets' ΔG, ID, smile from Modelseed database crawled by ../yeast-GEM/deltaG/code/get_seed_data.py
  4. ../yeast-GEM/deltaG/rxn_G.csv contains rxns' ΔG computed by https://github.com/maranasgroup/dGPredictor
  5. ../yeast-GEM/deltaG/code/adddeltaG.m adds ΔG into the model.
  6. ../yeast-GEM/deltaG/code/run_dgpredictor.py is used to calculate ΔG for the given rxns, but every mets in the rxn should have kegg identifer.

I hereby confirm that I have:

feiranl commented 1 year ago

As a suggestion for better tracking, it would be preferable to save the Excel file in either a .tsv or .csv format. By the way, may I know the way deltaG will be utilized in the future? Is this calculation straightforward or will it be automatically computed for any new reactions added to the model in the future? Additional comments will be provided later this week.

cheng-yu-zhang commented 1 year ago

@feiranl I have change xlsx to csv. And run_dgpredictor.py is able to calculate the deltaG of given rxns. Gibbs free energy is a thermodynamic potential that describes the reaction directionalities. On the basis of Gibbs-Helmhotz equation and Van 't Hoff equation, we can intergate metabolic concentration or metabolome into yeast-GEM. Because intracelluar metabolic concentration may vary with different growth period and the environment, especially temperature . Thus, the reaction directionalities may change. So, biochemically unrealistic solutions of FBA like free high-flux cycles or thermodynamically infeasible pathways can be avoided.

edkerk commented 1 year ago

Updated the process, now available from v8_6_4.m (in line with coming yeast-GEM version):

  1. Rebased to #304 to have model with latest reactions.
  2. Note that because of point 1, the dG-predictor has not yet been applied on all reactions. @cheng-yu-zhang model.rxns(4063:end) should be predicted with dG-predictor.
  3. The current code generates a model with the metDeltaG and rxnDeltaG fields, but these can currently not be exported in YAML or SBML model format. This can quite easily be implemented for YAML (to be done here: https://github.com/SysBioChalmers/RAVEN/issues/280), but uncertain on how to implement for SBML. This is the reason why the model files in this PR are not having any change included.
  4. loadYeastModel and saveYeastModel will add and save the fields, based on the table in /data/database/model_*DeltaG.csv.
edkerk commented 1 year ago

With the code on https://github.com/SysBioChalmers/RAVEN/pull/500 the deltaG fields are written to/read from the YAML file.