YORO: Pipeline for LTR-Retrotransposon domains detection and classification using Deep Neural Network YOLO inspired

Genomic Object Detection: An Improved Approach for Transposable Element Detection and Classification Using Convolutional Neural Networks

Installation

We highly recommend to use and install Python packages within an Anaconda environment. First, download the lastest version of YORO

git clone https://github.com/simonorozcoarias/YORO.git

Go to the YORO folder and find the file named "requirements.txt". Then, create and setup the environment:

conda env create --name YORO --file environment.yml
conda activate YORO

Testing

After successfully installing YORO, you can test it by using any genome. In this case, we will use the genome of Oryza Sativa:

conda activate YORO

Then, you must run the following command:

wget -O test/rice.fasta http://rice.uga.edu/pub/data/Eukaryotic_Projects/o_sativa/annotation_dbs/pseudomolecules/version_7.0/all.dir/all.con

python3 pipelineDomain.py -f test/rice.fasta -d temp

Finally compare your results in the path 'temp/output.tab' with the file in 'test/output.tab'. If you obtain similar (or also the same) results, congrats! the YORO pipeline is now installed and funcional.

Parameters

To obtain all parameters for pipeline execution, please run:

python3 pipelineDomain.py -h

All parameters that can be configurated are:

• -h or --help: show this help message and exit.
• -f FASTA_FILE or --file FASTA_FILE: Fasta file containing DNA sequences (required).
• -o OUTPUTNAME or --outputName name: Filename of the output file. Default: output.tab
• -d DIRECTORY or --directory DIRECTORY: Output Path to save output file. Default: Current directory.
• -p THREADS or --threads THREADS: Number of threads to be used by YORO. Default: all available threads.
• -t THRESHOLD or --threshold THRESHOLD: Threshold value for presence filter. Default: 0.6.
• -c CYCLES or --cycles CYCLES: Numbers of cycles of detections. Default: 1
• -w WINDOW or --window WINDOW: Window size for object detection and secuence splitting. Default: 50000
• -m MODELPATH or -modelpath MODELPATH: Path to models weights file. Default: models/ALL_DOMAINS_REDUNDANTYOLO_domain_v15.hdf5

Fast execution

To execute this pipeline you just need provide the fasta file for predictions and run as follows:

python3 pipelineDomain.py -f FASTA_FILE.fasta

if you wish a different name for output file, run as follows:

python3 pipelineDomain.py -f FASTA_FILE.fasta -o ModifidedOutput.tab

Pipeline Output

This tool produces one tabular file with this format:

|id	 |ProbabilityPresence	 |Start	 |Length	 |Class	 |ProbabilityClass

where:

• id: Secuence identifier from input fasta File
• ProbabilityPresence: probability of Transposable element presence
• Start: Predicted Initial position in the secuence
• Length: Predicted lenght of Transposable element
• Class: Type of Transposable element predicted
• ProbabilityClass: Probability of this class of transposable element.

Citation

If you use any material from this repository, please cite us as following:

Orozco-Arias, S., Lopez-Murillo, L. H., Piña, J. S., Valencia-Castrillon, E., Tabares-Soto, R., Castillo-Ossa, L., Isaza, G. & Guyot, R. (2023). Genomic object detection: An improved approach for transposable elements detection and classification using convolutional neural networks. Plos one, 18(9), e0291925.