Ensemble Enzyme Prediction Pipeline (E2P2)

The Ensemble Enzyme Prediction Pipeline (E2P2) annotates protein sequences with Enzyme Function classes comprised of full, four-part Enzyme Commission numbers and MetaCyc reaction identifiers. It is the enzyme annotation pipeline used to generate the species-specific metabolic databases at the Plant Metabolic Network since 2013. E2P2 systematically integrates results from two molecular function annotation algorithms using an ensemble classification scheme. For a given genome, all protein sequences are submitted as individual queries against the base-level annotation methods.

Due to PRIAM's end of development and availability, we've replaced it with DeepEC and moved the current E2P2 version to "v5". You can still download the previous version in the "v4" branch.

Getting Started

The following instuctions are for users to set up the E2P2 pipeline on a Unix machine and start running, developing and/or testing the pipeline.

Prerequisites

This pipeline is tested on Ubuntu, CentOS, macOS and should theoretically run on all Linux distributions.

• Python 3
• NCBI BLAST+
• Java 1.5 or above **Currently it has been known that PRIAM might not work with Java version 11, we recommend using version 8 instead.
• PRIAM_Search utility V2 **A copy is temperarily included due to the source website being down.
• DeepEC A fork of DeepEC is provided as a submodule, check "environment.yml" for prerequisites.

Installing

Download E2P2 from E2P2 at GitHub

Currently the E2P2 pipeline does not support white spaces or other illegal symbols in files paths.

git clone --recurse-submodules https://github.com/carnegie/E2P2.git

• If folder "deepec" is empty, run the following command in the 'E2P2' folder

git submodule update --init

Download Reference Protein Sequence Dataset (RPSD) from https://ftp.dpb.carnegiescience.edu/rpsd/

• Version release_2024_07_31 and up.

Unzip and extract RPSD data (different argument for tar depending on the file extensions)

tar -xzf blastdb.tar.gz
tar -xzf deepec.tar.gz
tar -xzf weights.tar.gz
tar -xzf maps.tar.gz

tar -xjf blastdb.tar.bz2
tar -xjf deepec.tar.bz2
tar -xjf weights.tar.bz2
tar -xjf maps.tar.bz2

# If you want to just use the required arguments with a different version of RPSD data, 
# replace the "maps" and "weights" folders under "data" in "E2P2" using the above folders 
# with the same name. And replace the files under "deepec/deepec/data" with the files in deepec.tar.gz.

Usage Example

config.ini

In the project's data folder is a "config_template.ini".

Users need copy the template to the root folder as "config.ini" and edit the environmental variables.

python3 e2p2.py [-h] --input /PATH/TO/Araport11_genes.201606.pep.repr.fasta -o PATH/TO/output.pf e2p2 --threshold 0.5

Required Arguments

--input INPUT_FILE, -i INPUT_FILE: Path to input protein sequences file

e2p2: subparser argument, used to separate pipeline arguments and classifier/ensemble arguments

Optional Arguments Before "e2p2"

-h, --help            show this help message and exit
--input INPUT_FILE, -i INPUT_FILE
                    Path to input protein sequences file
--protein_gene PROTEIN_GENE_PATH, -pg PROTEIN_GENE_PATH
                    Provide a protein to gene map. This can be used to generate a splice variant removed fasta file and output the final version of e2p2.
--remove_splice_variants, -rm
                    Argument flag to remove splice variants.
--output OUTPUT_PATH, -o OUTPUT_PATH
                    Path to output file. By Default would be in the same folder of the input.
--temp_folder TEMP_FOLDER, -tf TEMP_FOLDER
                    Specify the location of the temp folder. By default would be in the same directory of the output.
--log LOG_PATH, -l LOG_PATH
                    Specify the location of the log file. By default would be "runE2P2.log" in the temp folder.
--verbose {0,1}, -v {0,1}
                    Verbose level of log output. Default is 0.
                                0: only step information are logged
                                1: all information are logged

Optional Arguments After "e2p2"

-h, --help: Show help message and exit

--blastp BLASTP, -b BLASTP
                    Command of or path to BLAST+ "blastp".
--num_threads NUM_THREADS, -n NUM_THREADS
                    Number of threads to run "blastp".
--blast_db BLAST_DB, -bd BLAST_DB
                    Path to rpsd blast database name. For example, "/PATH/TO/FOLDER/rpsd.fa", where you can find the following files in
                    /PATH/TO/FOLDER:rpsd.fa.phr; rpsd.fa.pin; rpsd.fa.psq
--blast_e_value BLAST_E_VALUE, -be BLAST_E_VALUE
                    Blastp e-value cutoff
--blast_weight BLAST_WEIGHT, -bw BLAST_WEIGHT
                    Path to weight file for the blast classifier
--python_path PYTHON_PATH, -py PYTHON_PATH
                    Command of or path to "java".
--deepec_path DEEPEC_PATH, -dp DEEPEC_PATH
                    Path to "deepec.py".
--ec_to_ef_mapping_path EC_TO_EF_MAPPING_PATH, -ee EC_TO_EF_MAPPING_PATH
                    Path to mapping file from ECs to EFs
--threshold THRESHOLD, -t THRESHOLD
                    Threshold for voting results. Default is 0.5.

Additional information

• Input protein sequences should be in FASTA format.
• Headers of the FASTA file should begin with the sequence ID followed by a space or '|'. For example: >AT1G01010.1 | NAC domain containing protein 1 | Chr1:3760-5630 FORWARD LENGTH=429 | 201606

Authors

• Bo Xue - bxuecarnegie

Previous versions

• Chuan Wang - Chuan Wang
• Lee Chae

Institution

Plant Metabolic Network
Department of Plant Biology
Carnegie Institution for Science
Stanford, CA 94305

Acknowledgments

• Special Thanks to
  • Thomas Bernard - PRIAM
  • Ludo Cottret - lipme - Singularity Container For Previous Versions