Welcome to the KOMB repo! Get ready to KOMB through your (large) metagenomes to find regions of biological (functional or taxonomic) interest!
Current version of KOMB has to be installed from source. It has only been tested on Linux systems.
In order to install KOMB you will need to install several dependencies first. We recommend using conda
(Miniconda download) for managing KOMB dependencies. Below is an example of installing required tools with conda
.
- Create a new
conda
environment and make sure thatconda-forge
andbioconda
channels are enabled.
conda create --name komb-env python=3.9
conda config --add channels conda-forge
conda config --add channels bioconda
- Install dependencies available through
conda
conda install bwa-mem2
conda install seqkit
conda install igraph>=0.10.0
Note: we have upgraded KOMB to be compatible with newer version of the igraph library which means that the versions 0.8.3 and older no longer will work. We hope that igraph API will stay consitent going forward, but we have no way to ensure that.
3. Install rustup
and nightly
toolchain and build ggcat
from source
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
rustup toolchain install nightly
git clone https://github.com/algbio/ggcat --recursive
cd ggcat/
cargo install --path crates/cmdline/ --locked
- Clone the repo and build the source code
git clone https://gitlab.com/treangenlab/komb.git.
cd komb
./autogen.sh
./configure
make; make install
- Now you will have a
komb2
executable in thekomb/bin
directory and you should be able to runKOMB.py
in order to run the complete KOMB pipeline.
You can test KOMB by running the following command:
KOMB.py -i example_data/reads1.fastq -j example_data/reads2.fastq -k 51
Full set of parameters available in the KOMB pipeline is shown below and can be accessed by running python KOMB.py --help
.
usage: KOMB.py [-h] -i INPUT_READS1 -j INPUT_READS2 [-o OUTPUT_DIR] [--keep-alignments] [-e LOG_FILE] [--overwrite] -k KMER_SIZE [-t NUM_THREADS] [-l MIN_UNITIG_LENGTH] [-v VERBOSITY] [-c MIN_COUNT]
[-m GGCAT_MEMORY] [--eulertigs | --greedy-matchtigs | --pathtigs] [--min-seed-length MIN_SEED_LENGTH]
KOMB Analysis Pipeline example: python KOMB.py -i <read1.fq> -j <read2.fq> -k <k-mer size> -t <threads>
optional arguments:
-h, --help show this help message and exit
--eulertigs Generate Eulertigs instead of unitigs
--greedy-matchtigs Generate greedy matchtigs instead of unitigs
--pathtigs Generate pathtigs instead of unitigs
Input/Output:
-i INPUT_READS1, --input-reads1 INPUT_READS1
Path to the first sequencing reads file for paired-end data in FASTQ format
-j INPUT_READS2, --input-reads2 INPUT_READS2
Path to the second sequencing reads file for paired-end data in FASTQ format
-o OUTPUT_DIR, --output-dir OUTPUT_DIR
Path to the first sequencing reads file for paired-end data in FASTQ format
--keep-alignments Keep SAM files after the graph has been constructed (might require a lot of disk space)
-e LOG_FILE, --log-file LOG_FILE
File for logging [default: stdout]
--overwrite Delete <output-dir> and create a new one in case it exists
Common arguments:
-k KMER_SIZE, --kmer-size KMER_SIZE
k-mer size used for the *tig construction and subsequent analyses
use -1 to let KOMB automatically pick a value [default: -1]
-t NUM_THREADS, --num-threads NUM_THREADS
Maximum number of threads you want programs to use, note that some might use less than the amount specified
-l MIN_UNITIG_LENGTH, --min-unitig-length MIN_UNITIG_LENGTH
Minimum length of a unitig to be kept for the analysis.
Value -1 indicates setting this to the read length [default]
Value 0 would result in keeping all unitigs, and values > 0 will apply the filter
-v VERBOSITY, --verbosity VERBOSITY
Logging level: 0 (DEBUG), 1 (INFO), 2 (ERROR)
GGCAT *tig construction:
-c MIN_COUNT, --min-count MIN_COUNT
Minimum count required to keep a kmer [default: 2]
-m GGCAT_MEMORY, --ggcat-memory GGCAT_MEMORY
Maximum memory usage for GGCAT (GB) [default: 8]
BWA MEM parameters:
--min-seed-length MIN_SEED_LENGTH
Minimum seed length. Matches shorter than the value will be missed.