You may still use GATB-minia-pipeline in 2023 to assemble short reads. When doing so, I recommend a few tweaks: skip installation of bwa/BESST, and use options --no-scaffolding --no-error-correction. You will still get a fine an efficient multi-k assembly, without paired-end information that doesn't bring much contiguity gain in paired-end short reads anyway. -Rayan

GATB-Minia-Pipeline is a de novo assembly pipeline for Illumina data. It can assemble genomes and metagenomes.

It is multi-k, to aim for high contiguity. Similar software: MEGAHIT, metaSPAdes.

The pipeline consists of:

• Bloocoo (error correction)
• Minia 3 (contigs assembly) based on the BCALM2 (unitigs assembly) tool
• BESST (scaffolding)

• Linux 64 bits (for Minia binary)

• bwa (for BESST)

• Python >= 2.7 and < 3 for BESST:

  • BESST

BESST does not have a solid Python 3 support, hence only Pytohn 2 is supported.

See next section for a quick way to install them. Note: these Python modules are only needed for BESST. You can skip them if you do not plan on performing scaffolding.

python2 -m pip install --user BESST

git clone --recursive https://github.com/GATB/gatb-minia-pipeline

cd gatb-minia-pipeline ; make test

Command line arguments are similar to SPAdes.

Paired reads assembly:

./gatb -1 read_1.fastq -2 read_2.fastq

paired-end reads given in a single file:

./gatb --12 interleaved_reads.fastq

Unpaired reads:

./gatb -s single_reads.fastq

The final assembly is in:

assembly.fasta

All other files are intermediary.

More input options are available. Type ./gatb for extended usage.

Since the pipeline is multi-k, it is unnecessary to specify a kmer size.

Don't copy the ./gatb script to a bin folder it is meant to stay in that directory.

If you have trouble installing BESST, just give up and run ./gatb nevertheless. You can just skip the scaffolding step and the pipeline will still generate contigs.

If you want to persist compiling, read on.

Can't install BESST ? Try Conda (with Python 2) or Activestate

To contact an author directly: [email protected]