ML Predict: XGBoost for Small Molecules

Category Paths

Follow one of these paths in the Orion user interface, to find the floe.

  • Solution-based/Hit to Lead/Properties

  • Task-based/ADME & Tox Assessment

  • Solution-based/Hit to Lead/Properties/Model Building

Description

This floe predicts a property of small, drug-like molecules using a pretrained machine learning model.

It runs a Scikit-learn based XGBoost model for prediction. This user-provided model can be generated using the floes with preface ML Build. All floes with Model ID 0 is XGBoost so make sure you select those.

All models run on either 2D Fingerprints or Custom User Fingerprints based on what was provided during training. If Custom User Fingerprints were used, make sure to input them in the custom_feature parameter.

The floe is very cheap and quick, taking a few cents for the property prediction of 10 molecules.

Outputs: * Success Data: The prediction using XGBoost.

Promoted Parameters

Title in user interface (promoted name)

Inputs

Input Small Molecule(s) Dataset to predict property of (in): The dataset(s) to read records from

  • Required

  • Type: data_source

Input XGBoost Model (xgm): XGBoost Machine learning model to predict property.

  • Required

  • Type: data_source

Outputs

Output Property for XGBoost Regression (out): Output dataset to which to write.

  • Required

  • Type: dataset_out

  • Default: Successful XGBoost ML Regression Prediction

Failed Property for XGBoost based Regression (failed_out): Output dataset to which to write.

  • Required

  • Type: dataset_out

  • Default: Failed XGBoost ML Regression Prediction

Machine Learning Model Options

Custom Feature (cf): If model generated using floe ML Build: Regression Model with Tuner using Feature Input, use the same custom feature vector here for prediction.

  • Type: field_parameter