Scripts used in the manuscript "Evaluating and improving the representation of bacterial contents in long-read metagenome assemblies".
There is one ready-to-use set of scripts (bin merging), one demo script (k-mer spectra plot functions), and supporting scritps. Python needs to be >=3.6 for formatted string literals. Scripts were used in a python=3.10.5 environment.
Given a hifiasm-meta assembly & circle rescue results and genome binning from a traditional binner (e.g. metaBAT2), this script tries to merge the two without introducing too much duplication.
Script has three required inputs: 1) a tab-delimited plain text file
containing contig lengths and names, or an assembly graph with sequences
in gfa format, 2) circle rescue fasta file, and 3) traditional binner bins,
either a directory containing the bins (fasta/fastq/fna/...), or
a list of files (via bash expansion).
All files are read once.
See also -h.
The output is a single file or lines written to STDOUT, containing circle resuce and bin names. Together, they form the merged binning result.
Example of a run from scratch, as done in the manuscript:
# assembly
hifiasm-meta -t32 -o asm hifi.fq.gz 2>log_asm
# binning plus post-processing of bins, if necessary.
# Here we assume binner_bin will be the output that
# contain fasta files (suffix being ".fa"), each of which is a putative MAG.
# merge circle rescue and binner bins
python bin_merger.py -o result -x fa \
<(gfa2l.py asm.p_ctg.gfa) asm.rescue.fa binner_bin/
See yam for generating input file,
which is a 2D integer matrix of size (1024, 1024). Value x at row i and column j
means that there are x k-mers from the reads which appears exactly x times in reads
and y times in the assembly.
example/ provides a pair of plotting input and output.