Bug in def get_predictions(parameter, outfile):

[lollotala]

There is an iteration in the original definition that is looping while calling always the first row to estimate mean and SD of prediction.

Replace with following:

def get_predictions(parameter, outfile):
    df = pd.read_csv(outfile)
    pred_col = []
    pred_logcol = []
    pred_sd_totcol = []
    pred_sd_aleacol = []
    pred_sd_epicol = []
    
    for ii in  range(df.shape[0]):
        unit = 'mM'
        if parameter=='kcat':
            parameter_print = 'k_{cat}'
            parameter_print_log = 'log_{10}(k_{cat})'
            target_col = 'log10kcat_max'
            unit = 's^(-1)'
        elif parameter=='km':
            target_col = 'log10km_mean'
            parameter_print = 'K_{m}'
            parameter_print_log = 'log_{10}(K_{m})'
        else:
            target_col = 'log10ki_mean'
            parameter_print = 'K_{i}'
            parameter_print_log = 'log_{10}(K_{i})'
    
        unc_col = f'{target_col}_mve_uncal_var'
        model_cols = [col for col in df.columns if col.startswith(target_col) and 'model_' in col]
    
        unc = df[unc_col].iloc[ii]
    
        prediction = df[target_col].iloc[ii]
        prediction_linear = np.power(10, prediction)
    
        model_out = df[target_col].iloc[ii]
        model_outs = np.array([df[col].iloc[ii] for col in model_cols])
        # print(model_outs)
        epi_unc = np.var(model_outs)#np.sum(np.power(2, model_outs))/10. - np.power(2, model_out)
        alea_unc = unc - epi_unc
        epi_unc = np.sqrt(epi_unc)
        alea_unc = np.sqrt(alea_unc)
        unc = np.sqrt(unc)
        
        # print(unc-epi_unc-alea_unc)
        # def display_outs(prediction_type, out, alea_output, epi_output, unit):
        pred_col.append(prediction_linear)
        pred_logcol.append(prediction)
        pred_sd_totcol.append(unc)
        pred_sd_aleacol.append(alea_unc)
        pred_sd_epicol.append(epi_unc)

    df[f'Prediction_({unit})'] = pred_col
    df['Prediction_log10'] = pred_logcol
    df['SD_total'] = pred_sd_totcol
    df['SD_aleatoric'] = pred_sd_aleacol
    df['SD_epistemic'] = pred_sd_epicol

    return df

output_final = get_predictions(parameter, outfile)
output_final.to_csv(f'{outfile}')
print('Output saved to', outfile)