UniBioPAN is a versatile tool for the analysis and prediction of bioactive peptide activity. It provides a user-friendly interface and flexible options for both web-based and local execution.
The UniBioPAN web server allows for easy prediction of bioactive peptides:
- Sequence Input: Enter FASTA-formatted peptide sequences (up to 10).
- Threshold Selection: Adjust the threshold to control the prediction stringency.
- Activity Selection: Choose specific bioactivities to predict or "all activity" for a comprehensive analysis.
- Task Submission: Submit the task and wait for results to appear.
For users with limited resources or those who prefer a cloud-based solution, UniBioPAN integrates seamlessly with Google Colab.
- Download Files: Obtain the necessary files from the GitHub repository.
- Environment Setup: Execute provided code to install the required environment within Colab.
- Parameter Configuration: Adjust parameters as needed and train your model.
- Prediction: Use the best-performing model to predict bioactivity on your input data.
For maximum flexibility and local execution, UniBioPANscript offers a powerful Python interface.
- Environment Setup: Create a conda environment using the provided environment.yml file.
- Download Script: Obtain the UniBioPAN script files from GitHub.
- Command-Line Usage:
Use python unibiopan.py -h to display the help information.
Train a model:
> python unibiopan.py -t -tf Dataset/train.xlsx -ef Dataset/test.xlsx
Make predictions:
> python unibiopan.py -p -lp model/best_model.h5 -pf predict/input.xlsx
- Prediction Input: FASTA, CSV, or XLSX
- Training/Evaluation Input: XLSX (with 0 for positive and 1 for negative samples)
- Prediction Output: Table format, with 1 for predicted bioactive peptides, 0 for predicted inactive peptides, and 'x' for sequences that exceed the model's length limit or contain non-standard amino acids.
UniBioPAN predictions are based on computational models. Experimental validation is crucial to confirm the bioactivity of any peptide.