.
,'/ \`.
|\/___\/|
\'\ /`/ ██╗ █████╗ ███████╗ ██████╗ ███████╗██████╗
`.\ /,' ██║██╔══██╗██╔════╝██╔════╝ ██╔════╝██╔══██╗
| ██║███████║█████╗ ██║ ███╗█████╗ ██████╔╝
| ██ ██║██╔══██║██╔══╝ ██║ ██║██╔══╝ ██╔══██╗
|=| ╚█████╔╝██║ ██║███████╗╚██████╔╝███████╗██║ ██║
/\ ,|=|. /\ ╚════╝ ╚═╝ ╚═╝╚══════╝ ╚═════╝ ╚══════╝╚═╝ ╚═╝
,'`. \/ |=| \/ ,'`.
,' `.|\ `-' /|,' `.
,' .-._ \ `---' / _,-. `.
,' `-`-._,-'-' `.
'
Jaeger is a tool that utilizes homology-free machine learning to identify phage genome sequences that are hidden within metagenomes. It is capable of detecting both phages and prophages within metagenomic assemblies.
If you use Jaeger in your work, please consider citing its preprint:
- Jaeger: an accurate and fast deep-learning tool to detect bacteriophage sequences
Yasas Wijesekara, Ling-Yi Wu, Rick Beeloo, Piotr Rozwalak, Ernestina Hauptfeld, Swapnil P. Doijad, Bas E. Dutilh, Lars Kaderali
bioRxiv, 2024.09.24.612722
To cite the code itself:
- Jaeger: an accurate and fast deep-learning tool to detect bacteriophage sequences
The performance of the Jaeger workflow can be significantly increased by utilizing GPUs. To enable GPU support, the CUDA Toolkit and cuDNN library must be accessible to conda.
# create conda environment and install jaeger
mamba create -n jaeger -c bioconda jaeger-bio==1.2
# activate environment
conda activate jaegerTest the installation with test data
jaeger test# create a conda environment and activate
mamba create -n jaeger -c nvidia -c conda-forge cuda-nvcc "python>=3.11,<=3.12" pip
conda activate jaeger
# OR create a virtual environment using venv
python3 -m venv jaeger
source jaeger/bin/activate
# to install jaeger with GPU support
pip install jaeger-bio[gpu]
# to install without GPU support
pip install jaeger-bio[cpu]
# to install on a Mac(arm)
pip install jaeger-bio[darwin-arm]
# test the installation
jaeger test# create a conda environment and activate
mamba create -n jaeger -c nvidia -c conda-forge cuda-nvcc "python>=3.11,<3.12" pip
conda activate jaeger
# OR create a virtual environment using venv
python3 -m venv jaeger
source jaeger/bin/activate
# install jaeger
# to install with GPU support
pip install --no-cache-dir "jaeger-bio[gpu] @ git+https://github.com/MGXlab/Jaeger@dev"
# to install without GPU support
pip3 install --root-user-action=ignore --no-cache-dir "jaeger-bio[cpu] @ git+https://github.com/MGXlab/Jaeger@dev"
# to install on a Mac(arm)
pip3 install --root-user-action=ignore --no-cache-dir "jaeger-bio[darwin-arm] @ git+https://github.com/MGXlab/Jaeger@dev"
# test the installation
jaeger test
If you're using Apptainer on a cluster, it's recommended to build the container on your local machine and then transfer it to the cluster.
# get the container def
wget -O jaeger_singularity.def https://raw.githubusercontent.com/Yasas1994/Jaeger/dev/singularity/jaeger_singularity.def
# get the configuration file
wget -O config.json https://raw.githubusercontent.com/Yasas1994/Jaeger/dev/src/jaeger/data/config.json
# to build the container
apptainer build jaeger.sif singularity/jaeger_singularity.def
# test container
apptainer run --nv jaeger.sif jaeger --help
# test the installation
apptainer run --nv jaeger.sif jaeger test
# list jaeger models available for download
apptainer run --nv jaeger.sif download --list
# download jaeger models
apptainer run --nv jaeger.sif download --model jaeger_57341_1.5M_fragment --path /path/to/save/model --config /path/to/config.json
# run jaeger
apptainer run --nv jaeger.sif predict --model jaeger_57341_1.5M_fragment --config /path/to/config.json -i /path/to/input.fasta -o /path/to/save/results
Starting from version 1.2.0, users will need to download the new models separately after installing Jaeger. However, for backward compatibility, Jaeger will still include the old model by default.
Use the --list flag to print out all models available for download
jaeger download --listThen to download the model and add it to the model path run
jaeger download --path /path/to/store/models --model jaeger_38341_1.4MIf you decide to change the model path later, or if you have a dir witg newly trained/tuned models register the path
jaeger register-models --path /new/model/pathOnce the environment is properly set up, using Jaeger is straightforward. The program can accept both compressed and uncompressed .fasta files containing the contigs as input. It will output a table containing the predictions and various statistics calculated during runtime.
jaeger predict -i input_file.fasta -o output_dir --batch 128To run jaeger with singularity
apptainer run --nv jaeger.sif jaeger predict -i input_file.fasta -o output_dir --batch 128You can control the number of parallel computations using this parameter. By default it is set to 96. If you run into OOM errors, please consider setting the --bactch option to a lower value. for example 96 is good enough for a graphics card with 4 Gb of memory.
All predictions are summarized in a table located at output_dir/<input_file>_default.jaeger.tsv
┌───────────────────────────────────┬────────┬────────────┬─────────┬───┬─────────────┬────────────────┬──────────────────┬───────────────┐
│ contig_id ┆ length ┆ prediction ┆ entropy ┆ … ┆ Archaea_var ┆ window_summary ┆ terminal_repeats ┆ repeat_length │
╞═══════════════════════════════════╪════════╪════════════╪═════════╪═══╪═════════════╪════════════════╪══════════════════╪═══════════════╡
│ NODE_1109_length_9622_cov_23.163… ┆ 9622 ┆ Phage ┆ 0.43 ┆ … ┆ 0.143 ┆ 1V1n2V ┆ null ┆ null │
│ NODE_1181_length_9275_cov_26.864… ┆ 9275 ┆ Phage ┆ 0.327 ┆ … ┆ 0.504 ┆ 4V ┆ null ┆ null │
│ NODE_123_length_36569_cov_24.228… ┆ 36569 ┆ Phage ┆ 0.503 ┆ … ┆ 1.554 ┆ 9V1n7V ┆ null ┆ null │
│ NODE_149_length_32942_cov_23.754… ┆ 32942 ┆ Phage ┆ 0.458 ┆ … ┆ 3.229 ┆ 3V1n1n11V ┆ null ┆ null │
│ NODE_231_length_24276_cov_21.832… ┆ 24276 ┆ Phage ┆ 0.502 ┆ … ┆ 1.467 ┆ 1V1n3V1n5V ┆ null ┆ null │
└───────────────────────────────────┴────────┴────────────┴─────────┴───┴─────────────┴────────────────┴──────────────────┴───────────────┘
This table provides information about various contigs in a metagenomic assembly. Each row represents a single contig, and the columns provide information about the contig's ID, length, the number of windows identified as prokaryotic, viral, eukaryotic, and archaeal, the prediction of the contig (Phage or Non-phage), the score of the contig for each category (bacterial, viral, eukaryotic and archaeal), and a summary of the windows. The table can be used to identify potential phage sequences in the metagenomic assembly based on the prediction column. The score columns can be used to further evaluate the confidence of the prediction and the window summary column can be used to understand the count of windows that contributed to the final prediction.
jaeger run --help
## Jaeger 1.1.30 (yet AnothEr phaGe idEntifier) Deep-learning based bacteriophage discovery
https://github.com/Yasas1994/Jaeger.git
usage: jaeger run -i INPUT -o OUTPUT
options:
-h, --help show this help message and exit
-i INPUT, --input INPUT
path to input file
-o OUTPUT, --output OUTPUT
path to output directory
--fsize [FSIZE] length of the sliding window (value must be 2^n). default:2048
--stride [STRIDE] stride of the sliding window. default:2048 (stride==fsize)
-m {default,experimental_1,experimental_2}, --model {default,experimental_1,experimental_2}
select a deep-learning model to use. default:default
-p, --prophage extract and report prophage-like regions. default:False
-s [SENSITIVITY], --sensitivity [SENSITIVITY]
sensitivity of the prophage extraction algorithm (between 0 - 4). default: 1.5
--lc [LC] minimum contig length to run prophage extraction algorithm. default: 500000 bp
--rc [RC] minium reliability score required to accept predictions. default: 0.2
--pc [PC] minium phage score required to accept predictions. default: 3
--batch [BATCH] parallel batch size, set to a lower value if your gpu runs out of memory. default:96
--workers [WORKERS] number of threads to use. default:4
--getalllogits writes window-wise scores to a .npy file
--getsequences writes the putative phage sequences to a .fasta file
--cpu ignore available gpus and explicitly run jaeger on cpu. default: False
--physicalid [PHYSICALID]
sets the default gpu device id (for multi-gpu systems). default: 0
--getalllabels get predicted labels for Non-Viral contigs. default: False
-v, --verbose Verbosity level : -vvv warning, -vv info, -v debug, (default info)
Misc. Options:
-f, --overwrite Overwrite existing files
Jaeger can be integrated into python scripts using the jaegeraa python library as follows. currently the predict function accepts 4 different input types.
- Nucleotide sequence -> str
- List of Nucleotide sequences -> list(str,str,..)
- python file object -> (io.TextIOWrapper)
- python generator object that yields Nucleotide sequences as str (types.GeneratorType)
- Biopython Seq object
from jaegeraa.api import Predictions
model=Predictor()
predictions=model.predict(input,stride=2048,fragsize=2048,batch=100)
model.predict()returns a dictionary of lists in the following format
{'contig_id': ['seq_0', 'seq_1'],
'length': [19000, 10503],
'#num_prok_windows': [0, 0],
'#num_vir_windows': [9, 0],
'#num_fun_windows': [0, 5],
'#num_arch_windows': [0, 0],
'prediction': ['Phage', 'Non-phage'],
'bac_score': [-1.9552012549506292, -1.9441368103027343],
'vir_score': [6.6312947273254395, -3.097817325592041],
'fun_score': [-5.712721400790745, -0.6870137214660644],
'arch_score': [-2.4369852013058133, -0.8941479325294495],
'window_summary': ['9V', '5n']}
This dictionary can be easily converted to a pandas dataframe using DataFrame.from_dict() method
import pandas as pd
df = DataFrame.from_dict(predictions)- The program expects the input file to be in .fasta format.
- The program uses a sliding window approach to scan the input sequences, so the stride argument determines how far the window will move after each scan.
- The batch argument determines how many sequences will be processed in parallel.
- The program is compatible with both CPU and GPU. By default, it will run on the GPU, but if the --cpu option is provided, it will use the specified number of threads for inference.
- The program uses a pre-trained neural network model for phage genome prediction.
- The --getalllabels option will output predicted labels for Non-Viral contigs, which can be useful for further analysis. It's recommended to use the output of this program in conjunction with other methods for phage genome identification.
jaeger run -p -i NC_002695.fna -o outdir
The outdir will contain the following files
|____Escherichia_coli_O157-H7_prophages
| |____plots
| | |____NC_002695_Escherichia_coli_O157-H7_jaeger.pdf
| |____prophages_jaeger.tsv
|____Escherichia_coli_O157-H7_jaeger.log
|____Escherichia_coli_O157-H7_default_jaeger.tsv
users can find the following visulaization in the plots directory
list of prophage coordinates can be found in prophages_jaeger.tsv
┌─────────────┬────────────┬──────────┬──────────┬───┬──────────┬────────┬────────────┬────────────┐
│ contig_id ┆ alignment_ ┆ identiti ┆ identity ┆ … ┆ gc% ┆ reject ┆ attL ┆ attR │
│ ┆ length ┆ es ┆ ┆ ┆ ┆ ┆ ┆ │
╞═════════════╪════════════╪══════════╪══════════╪═══╪══════════╪════════╪════════════╪════════════╡
│ NC_002695 ┆ 16.0 ┆ 16.0 ┆ 1.0 ┆ … ┆ 0.435049 ┆ false ┆ GCACCATTTA ┆ GCACCATTTA │
│ Escherichia ┆ ┆ ┆ ┆ ┆ ┆ ┆ AATCAA ┆ AATCAA │
│ coli O157-… ┆ ┆ ┆ ┆ ┆ ┆ ┆ ┆ │
│ NC_002695 ┆ 15.0 ┆ 15.0 ┆ 1.0 ┆ … ┆ 0.493497 ┆ false ┆ GCTTTTTTAT ┆ GCTTTTTTAT │
│ Escherichia ┆ ┆ ┆ ┆ ┆ ┆ ┆ ACTAA ┆ ACTAA │
│ coli O157-… ┆ ┆ ┆ ┆ ┆ ┆ ┆ ┆ │
│ NC_002695 ┆ 60.0 ┆ 60.0 ┆ 1.0 ┆ … ┆ 0.511819 ┆ false ┆ TGGCGGAAGC ┆ TGGCGGAAGC │
│ Escherichia ┆ ┆ ┆ ┆ ┆ ┆ ┆ GCAGAGATTC ┆ GCAGAGATTC │
│ coli O157-… ┆ ┆ ┆ ┆ ┆ ┆ ┆ GAACTCTGGA ┆ GAACTCTGGA │
│ ┆ ┆ ┆ ┆ ┆ ┆ ┆ AC… ┆ AC… │
│ NC_002695 ┆ 16.0 ┆ 16.0 ┆ 1.0 ┆ … ┆ 0.499516 ┆ false ┆ TTCTTTATTA ┆ TTCTTTATTA │
│ Escherichia ┆ ┆ ┆ ┆ ┆ ┆ ┆ CCGGCG ┆ CCGGCG │
│ coli O157-… ┆ ┆ ┆ ┆ ┆ ┆ ┆ ┆ │
│ NC_002695 ┆ 14.0 ┆ 14.0 ┆ 1.0 ┆ … ┆ 0.529465 ┆ false ┆ CGTCATCAAG ┆ CGTCATCAAG │
│ Escherichia ┆ ┆ ┆ ┆ ┆ ┆ ┆ TGCA ┆ TGCA │
│ coli O157-… ┆ ┆ ┆ ┆ ┆ ┆ ┆ ┆ │
└─────────────┴────────────┴──────────┴──────────┴───┴──────────┴────────┴────────────┴────────────┘
You can use phage_contig_annotator to annotate and visualize Jaeger predictions.
This work was supported by the European Union’s Horizon 2020 research and innovation program, under the Marie Skłodowska-Curie Actions Innovative Training Networks grant agreement no. 955974 (VIROINF), the European Research Council (ERC) Consolidator grant 865694
The ascii art logo is from https://ascii.co.uk/