MetaCarvel is an updated version of previous metagenome scaffolder Bambus 2. To run MetaCarvel, you will need Python 3.7.x, Samtools, Bedtools, Networkx(Version >= 2.5), NumPy,and OGDF.
You can install Networkx as described here. MetCarvel can work with the latest NetworkX version 2.5 Briefly, you need to run following:
pip install numpy (tested with version 1.20)
pip install networkx>=2.5
The detailed documentation and tutorial to install and run MetaCarvel can be found on Wiki.
To run MetaCarvel, run the following;
python run.py -h
usage: run.py [-h] -a ASSEMBLY -m MAPPING -d DIR [-r REPEATS] [-k KEEP]
[-l LENGTH] [-b BSIZE] [-v VISUALIZATION]
MetaCarvel: A scaffolding tool for metagenomic assemblies
optional arguments:
-h, --help show this help message and exit
-a ASSEMBLY, --assembly ASSEMBLY
assembled contigs
-m MAPPING, --mapping MAPPING
mapping of read to contigs in bam format
-d DIR, --dir DIR output directory for results
-r REPEATS, --repeats REPEATS
To turn repeat detection on
-k KEEP, --keep KEEP Set this to keep temporary files in output directory
-l LENGTH, --length LENGTH
Minimum length of contigs to consider for scaffolding
in base pairs (bp)
-b BSIZE, --bsize BSIZE
Minimum mate pair support between contigs to consider
for scaffolding
-v VISUALIZATION, --visualization VISUALIZATION
To generate .db file for AsmViz visualization program
This will generate a bunch of files in the output directory. If you are interested in output of each step of the scaffolding process, these files can be useful. The final output files are scaffolds.fasta - which contains sequences of scaffolds and scaffolds.agp is an agp style information for assignment of contigs to scaffolds.
Please cite MetaCarvel as follows: Ghurye, J., Treangen, T., Fedarko, M., Hervey, W. J., & Pop, M. (2019). MetaCarvel: linking assembly graph motifs to biological variants. Genome biology, 20(1), 1-14.
NOTE: This tool is still under active development and may produce errors while running. Please report these as github issues so that we can fix them as we develop the software. For any questions, please email jayg@cs.umd.edu.