AlexanderKroll / kcat_prediction

MIT License
13 stars 5 forks source link

KeyError for 'DRFP' and 'KM' features #2

Open muhammad-faizan-122 opened 1 year ago

muhammad-faizan-122 commented 1 year ago

Hi @AlexanderKroll data/kcat_data/splits/train_df_kcat_with_KM_and_flux.pkl doesn't contain the 'DRFP' and 'KM' features. When I tried to load dataset using code given in last section of code/model_fitting/01 Training xgboost models with enzyme and reaction information.ipynb. Can you please help me in this?

AlexanderKroll commented 1 year ago

Hi Muhammmad,

thanks for pointing this out! I updated the zenodo data.zip file and it should now contain the updated files with the 'DRFP' and 'KM' features (see updated zenodo-link on Readme page).

muhammad-faizan-122 commented 1 year ago

Can you please give suggestion to further improve the performance of Deep learning model?

wang-ff commented 6 months ago

Hi @AlexanderKroll I also have this question just like Muhammmad. data/kcat_data/splits/train_df_kcat_with_KM_and_flux.pkl doesn't contain the 'DRFP' and 'KM' features. When I tried to load dataset using code given in last section of code/model_fitting/01 Training xgboost models with enzyme and reaction information.ipynb.I have already use the lastest zenodo data.zip(v3-8367052.zip) but still met this problem.Could you please help me in this.

AlexanderKroll commented 5 months ago

Hi Wang-ff,

I just downloaded the whole data folder from here https://zenodo.org/records/8367052 (v3) and when I open "data/kcat_data/splits/train_df_kcat_with_KM_and_flux.pkl", the last three columns are "KM", "flux", and "DRFP". Could you please check again that you can't find these columns?

Best, Alex