TnpB Model Setup and Usage Guide

Overview

This guide details the steps necessary for setting up the environment, training, and using the TnpB model for both training and inference.

Environment Setup

Creating a Virtual Environment

We recommend using Conda to manage your Python environments. Follow these steps to set up a dedicated environment for this project:

1. Create the Environment:

   conda create --name Tnpb 
   conda activate Tnpb
   conda install -c anaconda python=3.6
   pip install -r requirements.txt
   
   
   

Model Training

To train the TnpB model, follow these steps:

1. Data Preparation
   • Run data_preprocess.ipynb to process the data and save it in the /data/processed directory.
2. Model Training and Evaluation
   • Execute models_trainvaltest.ipynb or models_trainvaltest.py to train and evaluate the models.
   • Supported models include: FFN, RNN, CNN, Transformer.
   • Trained models are saved in the /output directory.
3. Inference
   • Use models_inference.ipynb or models_inference.py for running inference with trained models.
4. Sample Testing
   • Execute user_sample_inference.ipynb to test the model on your data. Example test file: data/Endogenous_spacers_TnpB_list.csv.
   • Predictions are saved in the /output directory.

Model Inference

For running model inference:

1. Setup
   • Ensure the environment is set up using requirements.txt.
   • Place your Excel data file containing target sequences at ./data/Endogenous_spacers_TnpB_list.csv.
2. Model Download
   • Download the trained model from this Dropbox link and add it to the /output directory.
3. Run Inference
   • Run user_sample_inference.ipynb to test your sample data. Predicted outputs are stored in the /output directory.
   • You can also directly run python user_sample_inference.py -data_name 'Endogenous_spacers_TnpB_list.csv'

Web Interface

• Access our web-based interface at tnpb.app to use the model directly without installing or running any Python code.

License

• By downloading or using the code, you are consenting to be and agree to be bound by all of the terms of this limited license described in license.txt file.