SemiBin2
SemiBin2 bins metagenomic contigs into metagenome-assembled genomes (MAGs) with deep learning, using the sequence composition and read coverage of the contigs. It works with assemblies from both short and long reads, and includes pretrained models for common environments such as the human gut, soil and ocean.
License
Free to use and open source under the MIT License.
Available
- Roihu-CPU: 2.6.0 (module
py-semibin), via thebio-appsmodule.
Usage
SemiBin2 is part of the bio-apps collection on Roihu. Load the bio-apps module tree and then the SemiBin2 module:
SemiBin2 needs the contigs and a sorted BAM file from mapping the reads to the contigs. Bin a single sample with a pretrained model:
SemiBin2 single_easy_bin -i contigs.fa -b sample.sorted.bam -o semibin_out \
--environment human_gut -t 16
The pretrained models are human_gut, dog_gut, ocean, soil, cat_gut,
human_oral, mouse_gut, pig_gut, built_environment, wastewater,
chicken_caecum and global. Leave out --environment to train a new model on
your own data instead, which takes longer. For contigs assembled from long reads,
add --sequencing-type long_read.
Always set the number of threads with -t. By default, SemiBin2 uses all the
cores of the node, not only the cores reserved for your job.
The bins are written to semibin_out/output_bins as gzipped FASTA files
(SemiBin_<number>.fa.gz), and the bin of each contig to
semibin_out/contig_bins.tsv. If a downstream tool needs uncompressed files, add
--compression none to write the bins as SemiBin_<number>.fa.
To bin several samples together, combine their contigs with
SemiBin2 concatenate_fasta, map the reads of each sample to the combined contigs
and run SemiBin2 multi_easy_bin. See the
SemiBin2 documentation for details.
To combine the bins of SemiBin2 with those of other binning tools, use MAGScoT.
Example batch script
#!/bin/bash
#SBATCH --job-name=semibin2
#SBATCH --account=<project>
#SBATCH --output=output_%j.txt
#SBATCH --error=errors_%j.txt
#SBATCH --partition=small
#SBATCH --time=08:00:00
#SBATCH --nodes=1
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=16
#SBATCH --mem=32G
module load bio-apps/v202603
module load py-semibin/2.6.0
SemiBin2 single_easy_bin -i contigs.fa -b sample.sorted.bam -o semibin_out \
--environment human_gut -t $SLURM_CPUS_PER_TASK
Replace <project> with your CSC project (for example project_2001234).
See creating a batch job script for Roihu for more information about running batch jobs.