TypeError: list indices must be integers or slices, not tuple when creating scatter plot of binary classification model

[michiel98fjhsad]

I have a binary classification problem for which I have developed a LightGBM classifier. I would like to plot global SHAP contributions for the X most important features, as in the scatter plot documentation the code shap.plots.scatter(shap_values[:, shap_values.abs.mean(0).argsort[-1]]) produces

My attempt is as follows:

clf_final = lightgbm.LGBMClassifier(**RSCV.best_estimator_.get_params())
clf_final.fit(X_train, y_train)

# compute SHAP values
explainer_LGB = shap.TreeExplainer(clf_final)
shap_values_LGB = explainer_LGB.shap_values(X_train)
shap.plots.scatter(shap_values_LGB[0:,"pet_mean_lag_t-6"])

But it produces the error message: TypeError: list indices must be integers or slices, not tuple

How could I get the desired plot for some of my features?