ushonium

usher + taxonium on Covid sequences.

Build

Build container with:

singularity build ushonium.img Singularity.def

Usage

The container has a script called ushonium, which makes a taxonium jsonl.gz file from fasta consensus sequences.

It uses mafft to align all sequences to the Covid reference (ignoring indels, so all the same length), makes an optimized tree with usher, and then uses taxoniumtools to make the taxonium jsonl file.

The filenames of consensus sequences need to be in a tab-delimited file with no headings and two columns:

Name of sample
FASTA file.

Assuming that TSV file is called samples.tsv, the usage is:

ushonium samples.tsv outdir

where the output directory outdir will be created. The final taxonium file is called 05.taxonium.jsonl.gz.

Use >1 CPU

Most of the time is spent making the MSA, which by default uses 1 cpu. Run in N in parallel using the option

--cpus N

This will hugely speed up the script.

Metadata

You can set the title that will appear in the taxonium browser with:

--title "My awesome tree"

You can also use a file of metadata for each sample (that has eg lineage, country etc). This needs to be tab-delimited with the first line having column headers. One column must have the name of the sample, and must exactly match the name given in samples.tsv. By default, this column is assumed to have the name strain, but you can change it to eg my_names with:

--metacol_name my_names

If the metadata file is called metadata.tsv, and we want columns col11 and col2, then use:

--metadata_tsv metadata.tsv --metacols col1,col2

Reference genome

The script ushonium was set up for covid, using the recommended genbank reference from taxoniumtools. This genbank file is included in the container, so there is no need to specify the reference genome when running ushonium unless you want to use a different reference. The option to change it to my_ref.gb is --ref_gb my_ref.gb.