Skip to content

seonwoo-min/TargetNet

Repository files navigation

TargetNet: Functional microRNA Target Prediction with Deep Neural Networks (Bioinformatics 2022)

Official Pytorch implementation of TargetNet | Paper

Abstract

Motivation: MicroRNAs (miRNAs) play pivotal roles in gene expression regulation by binding to target sites of messenger RNAs (mRNAs). While identifying functional targets of miRNAs is of utmost importance, their prediction remains a great challenge. Previous computational algorithms have major limitations. They use conservative candidate target site (CTS) selection criteria mainly focusing on canonical site types, rely on laborious and time-consuming manual feature extraction, and do not fully capitalize on the information underlying miRNA-CTS interactions.
Results: In this paper, we introduce TargetNet, a novel deep learning-based algorithm for functional miRNA target prediction. To address the limitations of previous approaches, TargetNet has three key components: (1) relaxed CTS selection criteria accommodating irregularities in the seed region, (2) a novel miRNA-CTS sequence encoding scheme incorporating extended seed region alignments, and (3) a deep residual network-based prediction model. The proposed model was trained with miRNA-CTS pair datasets and evaluated with miRNA-mRNA pair datasets. TargetNet advances the previous state-of-the-art algorithms used in functional miRNA target classification. Furthermore, it demonstrates great potential for distinguishing high-functional miRNA targets.


Installation

We recommend creating a conda environment from the TargetNet.yaml file as:

conda env create -f TargetNet.yaml

Alternatively, you can create a conda environment and install the necessary python packages as:

conda create -n TargetNet python=3.8
conda activate TargetNet
conda install pytorch==1.10.2 torchvision==0.11.3 cudatoolkit=11.3 -c pytorch -c conda-forge
pip install -r requirements.txt

Data Format

Extract the data.tar.gz file to obtain the dataset used for the manuscript as

tar -zxvf data.tar.gz

If you want to use other datasets, please follow the data format described as follows

  • The dataset must be in a tab-delimited file with at least 4 columns
  • The first row must be a header line (thus, will not be processed by the TargetNet algorithm).
  • The 1st ~ 4th columns must hold the following information
    • [1st column] miRNA id
    • [2nd column] miRNA sequence
    • [3rd column] mRNA id
    • [4th column] mRNA sequence -- mRNA sequence must be longer than 40 nucleotides
  • For the file containing train and validation datasets, it requires additional 2 columns
    • [5th column] label -- 0 or 1
    • [6th column] split -- train or val

Please refer to the provided dataset files for more detailed examples.

How to Run

Training a TargetNet model

You can use the train_model.py script with the necessary configuration files as

CUDA_VISIBLE_DEVICES=0 python train_model.py --data-config config/data/miRAW_train.json --model-config config/model/TargetNet.json --run-config config/run/run.json --output-path results/TargetNet_training/

The script will generate a TargetNet.pt file containing a trained model.
For using other datasets, modify the data paths specified in the miRAW_train.json data-config file.

Evaluating a TargetNet model

You can use the evaluate_model.py script with the necessary configuration files as

CUDA_VISIBLE_DEVICES=0 python evaluate_model.py --data-config config/data/miRAW_eval.json --model-config config/model/TargetNet.json --run-config config/run/run.json --checkpoint pretrained_models/TargetNet.pt --output-path results/TargetNet-evaluation/

The script will generate a tab-delimited *_outputs.txt file described as follows

  • The output file contains a TargetNet prediction for each miRNA-mRNA set from the given dataset.
    • [1st column] set_idx -- indicates the set index from the given dataset.
    • [2nd column] output -- TargetNet prediction score ranging from 0 to 1.
  • For binary classification results, use a threshold of 0.5 to binarize the output scores.

For using other datasets, modify the data paths specified in the miRAW_eval.json data-config file.


About

Official Pytorch implementation of TargetNet, Bioinformatics 2022

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages