nibtehaz/GO2Sum

GO2Sum is a deep learning based summarizer that generates human-readable summaries for GO term annotations made by protein function prediction methods.

GO2Sum

GO2Sum is a deep learning based summarizer that generates human-readable summaries for GO term annotations made by protein function prediction methods.

Copyright (C) 2023 Swagarika Jaharlal Giri, Nabil Ibtehaz, Daisuke Kihara, and Purdue University.

License: GPL v3. (If you are interested in a different license, for example, for commercial use, please contact us.)

Contact: Daisuke Kihara (dkihara@purdue.edu)

For technical problems or questions, please reach to Swagarika J Giri (giri2@purdue.edu).

Pre-required software

Python 3.9 : https://www.python.org/downloads/

Installation

1. Install git

2. Clone the repository in your computer

git clone https://github.com/kiharalab/GO2Sum.git && cd GO2Sum

3. Create a new environment and activate it

Install anaconda

3.1 install conda.

3.2 create a new environment go2sum and activate it

conda create --name go2sum python=3.9
conda activate go2sum

Each time when you want to run this code, simply activate the environment by

conda activate go2sum
conda deactivate    (If you want to exit) 

4. Build dependencies.

GO2SUM is based on simpleT5, a Pytorch-lightning implementation of T5 model

Install with pip and python.

pip install --upgrade simplet5

5. Prepare Data

Please download and unzip the models.zip files

https://kiharalab.org/GO2Sum/models.zip

6. Usage

Use GO2Sum to generate summary from annotated goterms

python3 main.py

            --input_file  ('Tab-separated file with Protein ID and GO Annotation list')

            --summary_type' ('Type of summary to generate :function, subunit, pathway, or all)

            --output_file ('Name of result file')

To use the script, follow these steps

6.1 Create a tab-separated input_file in the following format

Protein	GO_IDs
Q5AK66	GO:0000139;GO:0004609;GO:0005795;GO:0006646;GO:0006656;GO:0006659;GO:0010008;GO:0016540;GO:0036170;GO:0036171;GO:0036180;GO:0046872
A9AJN2	GO:0004609;GO:0005886;GO:0006646

Each row should correspond to a single protein and its associated GO IDs. The first column should contain the protein name, and the second column should contain a semicolon-separated list of GO IDs.

Sample input files named test1.tab and test.tab is provided in the test/ directory for your reference.

6.2 Run the script using the following command

python3 main.py 

                --input_file <path_to_input_file> 

                --summary_type <summary_type> 

                --output_file <output_file_suffix>

Replace <path_to_input_file> with the path to your input file, <summary_type> with one of the following options: function, pathway, subunit, or all and <output_file_suffix> with the suffix you want to use for the output file.

6.3 Example

For example, if your input file is named test1.tab, and you want to generate a summary of the function for each protein and save the output file with the suffix example, you would run the following command:

Example 1

python3 main.py --input_file test1.tab --summary_type function --output_file example.tab

This would generate the output files function_example.tab in the result directory.

Example 2

python3 main.py --input_file test1.tab --summary_type all --output_file example.tab

This would generate the output files function_example.tab, subunit_example.tab, and pathway_example.tab in the result directory.

Experiments

All the codes to run the experiments are provided in /Experiments directory

Citation:

"GO2Sum: Generating Human readable functional summary from GO terms, Swagarika Jaharlal Giri, Nabil Ibtehaz, and Daisuke Kihara, In submission (2023)"