ProteinScape: Protein Structure Prediction Application

ProteinScape is a web-based application designed to predict and visualize the 3D structure of proteins from amino acid sequences. The app utilizes the ESMFold API to provide accurate structural predictions and includes an interactive interface for researchers, educators, and students to explore protein structures and their features.

Features

• Single Protein Structure Prediction: Predict the 3D structure of a single protein sequence.
• Multiple Protein Structure Prediction: Predict the structure for multiple protein sequences (from FASTA file or manual input).
• 3D Visualization: Interactive 3D visualization of predicted protein structures using py3Dmol.
• Amino Acid Distribution Analysis: Graphical display of the amino acid composition in the input sequence.
• Sample Ramachandran Plot: A sample Ramachandran plot showing the phi (ϕ) and psi (ψ) angles for each sequence.

Tech Stack

• Frontend: Streamlit
• Backend: ESMFold API
• Visualization: py3Dmol
• Plotting: Matplotlib, Seaborn
• Data Handling: Pandas, Collections
• Requests: Python requests library for API calls

Installation

1. Clone the repository:

   git clone https://github.com/your-username/proteinscape.git
   cd proteinscape

2. Install the required dependencies:

   pip install -r requirements.txt

   Ensure that the following Python packages are installed:

   • streamlit
   • py3Dmol
   • matplotlib
   • seaborn
   • pandas
   • requests
   • numpy

3. Run the application:

   streamlit run app.py

   Open the provided localhost link in your browser.

Usage

Single Protein Structure Prediction

1. Enter a valid protein sequence (amino acids: A, C, D, E, F, G, H, I, K, L, M, N, P, Q, R, S, T, V, W, Y).
2. Click the "Predict Structure" button.
3. View the predicted 3D structure of the protein in the interactive viewer.
4. Download the predicted structure as a .pdb file.
5. View the amino acid distribution and a sample Ramachandran plot for further insights.

Multiple Protein Structure Prediction

1. Upload a FASTA file with multiple protein sequences or manually input sequences in FASTA format.
2. The application will predict the structure for each sequence and allow visualization and downloading of each result.
3. Similar to single predictions, view the amino acid distribution and Ramachandran plot for each sequence.

3D Visualization

• Use the py3Dmol viewer to interact with the predicted protein structure.
• Select different visualization styles: cartoon or stick.

API Information

The application uses the ESMFold API for protein structure prediction. The predicted output is displayed in .pdb format and is rendered in the web application for real-time interaction.

Example

Input

A valid protein sequence:

MKTAYIAKQRQISFVKSHFSRQDILDLWQYFSYGRAL

Output

• 3D predicted structure of the protein.
• Downloadable .pdb file.
• Bar chart of amino acid distribution.
• Sample Ramachandran plot.

Known Limitations

• The application can only predict structures for sequences shorter than 1,500 amino acids due to the ESMFold API limitations.
• The predictions are based on AI models and should not be used for clinical purposes.

License

This project is licensed under the MIT License. See the LICENSE file for more details.

Contact

For questions, feedback, or issues, please contact:

• Abdul Rehman Ikram hanzo7n@gmail.com