PFGPred - Plant Fusion Gene Predictor

Step 1: Preprocess Your Training Data

Before uploading, please preprocess your fusion data using the official PFGPred preprocessing and training utilities available on GitHub. These scripts will help you construct a labeled training table (for example, final_with_labels.csv with a label column of 0/1).

🔗 PFGPred GitHub Repository (Preprocessing & Training Guide)

After preprocessing, upload your labeled .CSV file here. Both training and optional test datasets must include a header row and must not exceed 10,000 rows each.

Caution: The training process is computationally intensive and may take several minutes to complete, depending on the size of your dataset and the server load.

Step 2: Upload Training Data

Step 3: Configure Training

Select Target Column

Ensure this column contains only '1' (positive) and '0' (negative) values.

Step 4: Set Model Parameters

The default, pre-optimized parameters will be used. This is recommended for most users.

Modify the hyperparameters for the models below.

XGBoost

Learning Rate (eta):

Max Depth:

Random Forest

Num. Estimators:

Max Depth:

LSTM

Epochs:

Batch Size:

PFGPred (Meta-Learner)

C (Regularization):

Step 5: Upload Test Data or Split Training Data

You can provide a separate test set to evaluate the model. If omitted, a portion of your training data will be used for validation based on the ratio you select below.

Select Train/Test Split Ratio

Training in Progress...