Cytopathic Effect Prediction App

Project Overview

This project is a machine learning application built to predict the cytopathic effect based on its SMILES representation. The machine learning model used is a RandomForestClassifier, which has been trained on molecular fingerprint data derived from SMILES strings. The project includes both model training and a user-friendly Streamlit app to make predictions on new molecules and visualize the model's performance metrics.

For more details, visit the GitHub Repository.

Technologies Used

• RDKit: To convert molecular SMILES strings into fingerprints.
• Scikit-learn: To train the Random Forest model and evaluate performance metrics.
• Imbalanced-learn: To balance the dataset using SMOTE (Synthetic Minority Over-sampling Technique).
• Joblib: To save and load the model and scaler.
• Streamlit: To build an interactive web app for predictions and model performance visualization.

Model Training Process

1. Data Preparation:

   • The SMILES strings are converted into molecular fingerprints using RDKit.
   • The dataset is balanced using SMOTE to handle any class imbalance.
   • The data is split into training and test sets.
   • The training data is scaled using StandardScaler.

2. Model Training:

   • A RandomForestClassifier is trained on the scaled training data.
   • The model and the scaler are saved using Joblib for future use.

3. Model Evaluation:

   • The model's accuracy is calculated on the test set.
   • A classification report is generated to summarize performance metrics such as precision, recall, and F1-score.

Streamlit App Features

The Streamlit app provides the following functionalities:

1. Single Molecule Prediction:

• Input: A SMILES string representing a molecule (e.g., "CCO" for ethanol).
• Output: The prediction of whether the molecule has a cytopathic effect (active or inactive) and the probability score of the prediction.

2. Model Performance Metrics:

The app visualizes key performance metrics based on test data. This includes:

• Confusion Matrix: A matrix displaying the number of true positive, true negative, false positive, and false negative predictions.
• ROC Curve: A graph showing the trade-off between sensitivity (true positive rate) and specificity (false positive rate). The ROC-AUC score is also displayed.
• Prediction Distribution: A bar chart showing the distribution of predicted classes (active/inactive).

Input and Output

Input

• Single Molecule Prediction: The app requires a SMILES string as input in the text box. Example:
  • C1=CC=CC=C1 (Benzene)
  • CCO (Ethanol)

Output

• The output of the app includes:
  • A prediction of whether the molecule has a cytopathic effect (active or inactive).
  • The probability associated with the prediction (e.g., 0.75 for a 75% chance that the molecule is active).
  • Visualizations such as confusion matrix, ROC curve, and prediction distribution.

How to Run

1. Install dependencies: Run the following commands to install the necessary packages:

   pip install rdkit-pypi scikit-learn joblib imbalanced-learn streamlit