Skip to content

Latest commit

 

History

History
81 lines (56 loc) · 3.23 KB

README.md

File metadata and controls

81 lines (56 loc) · 3.23 KB
Raw

CZ ID PhyloTree-NG workflow

This is the CZ ID PhyloTree-NG (next generation) workflow. Given a set of CZ ID samples and a taxon ID of the organism of interest, the workflow computes the relative genomic distance between samples and generates phylogenies of the samples. Each sample is represented by contigs (assembled reads) from the original input sample mapped to reference sequences for the organism of interest.

Depending on coverage and phylogenetic diversity, the resulting phylogenies range in precision from a heatmap (pairwise distance matrix) to a phylogram.

The workflow uses SKA split k-mers to create clusters and phylogenetic trees. This works for complete genomes as well as raw sequences (currently supporting only .fasta inputs).

Implementation details

Current workflow overview

The workflow takes as input the samples (each represented by sample name, mngs workflow run ID, contig fasta file, contig summary file) and reference accessions (idetified via a heuristic). The workflow then prepares FASTA files for each sample and reference:

  • For samples, we scan their merged hit summaries to determine which contigs map to taxa under the given reference taxon ID, then subset the sample's contigs FASTA file.

  • For references, we download their FASTA files from the AWS NCBI BLAST database S3 bucket using taxoniq.

We then feed the FASTA files to SKA to create kmer profiles and generate phylogenies.

Detailed workflow description

A detailed outline of the workflow steps is available in Phylotree-ng.md.

Old workflow

An in-depth description of the first generation PhyloTree pipeline from idseq-dag (in IDseq until August 2021) is available in Phylotree.md. The authoritative copies of the associated idseq-dag steps are in prepare_taxon_fasta.py and enerate_phylo_tree.py.

Running the workflow locally

First, clone the repo

git clone https://github.com/chanzuckerberg/czid-workflows.git

Then, build the docker image:

docker build -t phylotree-ng phylotree-ng

We then use the docker image to run the pipeline as follows:

miniwdl run --verbose phylotree-ng/run.wdl --input phylotree-ng/test/test_inputs.json

The inputs.json file will have the following structure:

{"samples":[

], "reference_taxon_id":64320,
"superkingdom_name":"viruses",
"additional_reference_accession_ids":["NC_012532.1", "NC_035889.1"],
"docker_image_id":"phylo-ng"}

# where the "samples" array contains instances of the following format
{"workflow_run_id":1, "sample_name":"X", "combined_contig_summary":".json", "contig_fasta":".fasta"}

The /analysis_support directory contains scripts that have been used to support the experimentation and validation of the phylotree pipeline.

Reference files

Filename Provenance
test/full_zika.tar.gz https://github.com/katrinakalantar/clustered-phylotree/blob/main/test/full_zika.tar.gz?raw=true