Too many charge unbalaced reactions

[AnnyMais03]

My model has more than 400 unbalanced reactions and I dont know why is this, trying to solve them manually takes an enormous amount of time, is there a way to have a script that directly takes the charges from BiGG database and then fixes the rest?

[AzizBenA]

def get_bigg_models_for_imbalanced_reactions(model, limit=None, verbose=True):
    """
    Retrieves BiGG model IDs for reactions that are mass/charge imbalanced in the given model.

    Parameters:
    - model: COBRApy model object
    - limit: optional int, max number of reactions to check (for testing or speed)
    - verbose: bool, whether to print progress

    Returns:
    - dict: mapping of reaction ID → model ID from BiGG database
    """
    imbalanced = check_mass_balance(model)
    reaction_model_mapping = {}

    for i, (reaction, balance_dict) in enumerate(imbalanced.items()):
        if limit and i >= limit:
            break
        reaction_id = reaction.id
        url = f'http://bigg.ucsd.edu/api/v2/universal/reactions/{reaction_id}'

        response = requests.get(url)
        if response.status_code == 200:
            data = response.json()
            models = data.get('models_containing_reaction', [])
            if models:
                reaction_model_mapping[reaction_id] = models[0]['bigg_id']
                if verbose:
                    print(f"{reaction_id} → {models[0]['bigg_id']}")
            else:
                if verbose:
                    print(f"{reaction_id} found, but no models listed.")
        else:
            print(f"Failed to retrieve {reaction_id}. Status code: {response.status_code}")

    print(f"\n✅ Retrieved model info for {len(reaction_model_mapping)} imbalanced reactions.")
    return reaction_model_mapping

def update_metabolite_charges_from_bigg(model, reaction_model_mapping):
    """
    Updates zero-charge metabolites in reactions using charge values from the BiGG database.

    Parameters:
    - model: COBRApy model object
    - reaction_model_mapping: dict mapping reaction IDs to BiGG model names

    Returns:
    - int: number of metabolites whose charges were updated
    """
    updated_count = 0

    for reaction_id, bigg_model_name in reaction_model_mapping.items():
        try:
            reaction = model.reactions.get_by_id(reaction_id)
        except KeyError:
            print(f"Reaction '{reaction_id}' not found in model.")
            continue

        for metabolite, coefficient in reaction.metabolites.items():
            if metabolite.charge == 0:
                url = f'http://bigg.ucsd.edu/api/v2/models/{bigg_model_name}/metabolites/{metabolite.id}'
                response = requests.get(url)

                if response.status_code == 200:
                    data = response.json()
                    bigg_charge = data.get('charge')
                    if bigg_charge is not None and bigg_charge != metabolite.charge:
                        print(f"Updating metabolite '{metabolite.id}'")
                        print(f" - Current charge: {metabolite.charge}")
                        print(f" - BiGG charge   : {bigg_charge}")
                        metabolite.charge = bigg_charge
                        updated_count += 1
                else:
                    print(f"Failed to retrieve data for '{metabolite.id}'. Status code: {response.status_code}")

    print(f"\nTotal metabolites updated: {updated_count}")
    return updated_count

Try to check this functions

[Gscorreia89]

@AzizBenA Am I correct that this is a fix to this issue #174 ?

[AzizBenA]

i think so !

[Gscorreia89]

OK, from what I noticed this is an issue with the universe models for bacteria, gram_pos, gram_neg. I think it would be best to apply this fix to the universe .xml models and fix the issue at root.