ScaRa is our scaffolding tool that uses come concepts from the HERA scaffolder paper (https://www.biorxiv.org/content/early/2018/06/13/345983). This repository contains both Python and C++ implementations.
- Clone the repository, and include all submodules.
git clone --recursive https://github.com/kkrizanovic/ScaRa.git
Note: if you omitted <--recursive>
from <git clone>
, run <git submodule update --init --recursive>
before proceeding.
After cloning, Python scripts are ready to go.
- Building C++ version
mkdir build
cd build
cmake ..
make
Python scripts require PYthon2.7. C++ version requires CMake 3.5.
ScaRa is run using <bridger.py>
script from the root folder of the repository. Running the script without arguments will print out a help message. The script should be run wtith the first parameter set to scaffold
in the following way:
bridger.py scaffold <contigs FASTA> <reads FASTA> <reads-contigs overlaps PAF/SAM> <reads-reads overlaps PAF/SAM options
options:"
-o (--output) <file> : output file to which the report will be written
C++ executable should be in the build folder. Running it without arguments will print out following help message.
How to run ScaRa :
ScaRa -f (--folder) <Input folder>
or
ScaRa -r <Reads file> -c <Contigs file> -o <Reads to Contigs Overlaps file> -s <Reads to Reds Overlaps file>
The program will perform one iteration of the algorithm
and output contigs to the standard output!.
<Input foder> must contain the following files:
- reads.fastq - reads in FASTQ/FASTA format
- readsToContigs.paf - overlaps between reads and contigs
- readsToReads.paf - overlaps between reads
- contigs.fasta - contigs in FASTA format
If Reads file, Contigs file or Overlaps are specified,
they will not be looked for in the Input folder!
Options:
-f (--folder) specify input folder for ScaRa
-r (--reads) specify reads file for ScaRa
-c (--contigs) specify contigs file ScaRa
-o (--overlapsRC) specify contig-read overlaps file for ScaRa
-s (--overlapsRR) specify read self overlaps file for ScaRa
-m (--multithreading) use multithreading
-v (--version) print program version
-h (--help) print this help message
Old ScaRa README (containing only the Python version) can be found at OLD ScaRa