NO prdicttion

[wangtao-cell]

Hello, Dr. Alexander Kroll.：
After I downloaded all the files and configured my environment, I can run the example you gave me, but there is no Prediction in the results, and the GNN FP and enzyme rep are empty.

[wangtao-cell]

GNN.py is empty !

[jackychen08]

I also ran into this issue but this is not actually due to GNN.py being empty. Instead, the error comes from line 58 in ES_predictions.py

if isinstance(gnn_rep, str) or isinstance(esm1b_rep, str) or isinstance(df_ES["enzyme"][ind], str):
			df_ES["complete"][ind] = False

This marks the corresponding entry as incomplete if the enzyme stored at that index is a string. This is counter-intuitive as the enzyme column is supposed to be strings as that is how the argument is passed to the function. It is also defined earlier in the code in line 21,26

enzyme_list = ["" if pd.isnull(value) else value for value in enzyme_list]
df_ES = pd.DataFrame(data = {"substrate" : substrate_list, "enzyme" : enzyme_list, "index" : list(range(len(substrate_list)))})

As a solution, swapping line 58 to be

if isinstance(gnn_rep, str) or isinstance(esm1b_rep, str) or  df_ES["enzyme"][ind] == "":
            df_ES["complete"][ind] = False

fixes the problem. This way, we keep marking entries as complete when they are strings, but mark them as incomplete if they have an empty field

[wangtao-cell]

Thank you very much, your explanation solved my problem. The issue was indeed in that area.