DeepIRES

DeepIRES: a hybird deep learning model for indentifying internal ribosome entry site in mRNA

EXPLANATION

This repository contains four folders: model, dataset, weights, data, and result.

Model

This folder contains python code files for constructing model.

Dataset

This folder contains orginal data, traing dataet and testing dataset we constructed.

Weights

This folder saves the model weights we trained.

Data

You can put your input data in .fa format in this folder to run prediction program. It already contains our testing sets

Result

This folder is used to save prediction output file

Installation of DeepIRES and environment

Download the repository and create corresponding environment.

git clone https://github.com/SongLab-at-NUAA/DeepIRES.git
cd ./DeepIRES
conda env create -f environment.yml

Then activate virtual enviroment

conda activate DeepIRES

USAGE

Run DeepIRES.py file to predict IRES

python DeepIRES.py -i input_file -o output_name 

-i input file : the input file in .fa or .fasta format.This file should be located in data folder containing the sequneces you want to predict.
-o output name : the output name you want to use.The result is saved in .csv format in result folder.

For example

python DeepIRES.py -i main_independent.fa -o main
python DeepIRES.py -i core.fa -o core
python DeepIRES.py -i 5utr_independent.fa -o 5utr

There are four columns in the prediction results tabular：sequence name, IRES score, the start locations of region may contain IRES, the termination locations of region may contain IRES.