ProTaxoVis 0.0.3

Map Blast results on a common-knowledge taxonomix (phylogenetic) tree

ProTaxoVis

This package contains tools to map BLAST results on the NCBI taxonomy phylogenetic (taxonomic) tree. It helps to analyse the presence/absence of proteins or genes in various taxa. This information can be useful, for example, for the analysis of two competing pathways (see e.g. Bockwoldt et al. 2019).

If you used this tool in your work, please cite:

TODO: Citation of our paper.

Installation

You will need Python 3.6+ and, optionally, BLAST+. Everything else can be install using pip. We suggest to use a virtual environment or Conda environment to install everything into. The examples below are using virtual environments with conda environment equivalents given in the comments.

# Create a virtual environment. Not strictly necessary,
# but generally recommended to avoid problems with conflicting versions
python3 -m venv taxoenv
# conda create -n taxoenv

# Activate the environment. You have to do this every time, you want
# to work with ProTaxoVis and open a new shell.
source /path/to/taxotest/taxoenv/bin/activate
# conda activate taxoenv

# Install ProTaxoVis
pip install protaxovis

# You need to tell taxfinder once to download the latest NCBI
# taxonomy information
taxfinder_update

# Now, you can run all the good stuff. Start, for example, with:
taxovis --init

After the installation, when opening a new terminal, you only have to reactivate the virtual/conda environment

source /path/to/taxotest/taxoenv/bin/activate
# conda activate taxoenv

Components

This package consists of four command line tools and one importable Python module. All command line tools can be run with --help to get help.

Workflow

Example data to run with can be downloaded from https://github.com/MolecularBioinformatics/ProTaxoVis-examples This data includes the results of a blast run against nr and has config files ready, allowing you to skip steps 2, 3, and 5 of the workflow below.

A typical workflow is described in (TODO: Our paper) and can be summed up to these steps:

1. Create a folder for your project. Open a terminal/shell in this folder and run taxovis --init.
2. Collect fasta files of your proteins or genes of interest. These are called seeds. Fasta files of multiple species for the same protein or gene are possible. Place the fasta files in the fasta folder of your project. Modify the files proteinlist.txt and limits.txt.
3. Run BLAST against the database of your choice (e.g. nr for proteins, nt for genes). This can be done using the command line tool or the NCBI BLAST website. It is important that taxonomy ids are included in the results. This is given in all NCBI databases. For your own databases, please consult the documentation for makeblastdb on how to include taxonomy ids. If you run BLAST on the NCBI Website, make sure to download the result as "Single-file XML2" and save them to the blastresults folder with the same filename as the corresponding fasta file but with .xml instead of .fasta.
4. Run the command taxovis --all.
5. Modify the files tree_config.txt and tree_to_prune.txt.
6. Run the command taxotree --show. You may want to go forth and back between modifying tree_to_prune.txt and visualizing the tree until the focus of the tree is right.
7. Save the tree using taxotree --outfile tree_name.pdf.
8. If you are interested in the interactive heatmap, modify heatmap_config.txt and run taxovis --only intheat.

Output files

histograms/ shows the length distribution of BLAST results for each seed. blastmappings/ shows where the seeds map on the BLAST results dependent on the e-value cutoff. Both can be used to tweak limits.txt.

trees/ contains the trees for each seed. These trees contain all species in which the seed was found by BLAST. general.tre is the combined tree for all seeds.

heatmap.html is an interactive heatmap of selected taxa. You can select taxa and other parameters in heatmap_config.txt and re-run taxovis --only intheat.

matrix.csv is a table that shows how similar two seeds are. The number of the cell where seed A and seed B cross is the e-value at which the BLAST search of seed A found seed B.

Other command line tools

blast2fasta can be used to download sequences based on BLAST results.

Importable modules

taxovis can not only used as command line tool, but also imported for your own workflows without the hardcoded filenames etc.

venn can create Venn diagrams. There is no direct integration with the other tools in this package, but it may serve useful for your own custom workflows.

sample_taxids can be used to randomly sample taxids from a file of taxids, and write these out to a heatmap configuration file. Rerunning the heatmap step of taxovis will then redraw the heatmap with sampled taxids.

Utilities

There are some utility scripts that are only on Github and not downloaded by pip. You can find them in the utils folder in the repository. These scripts are not polished and are meant to be changed before using them. They might come in handy, though, so we did not delete them outright. Each of the scripts has a little information about them in the top of the file.

lineage_value.py gives an overview about how present given seeds are along a phylogenetic lineage. If you, for example, given human (taxonomy id 9606) as target, it will show, how well the seeds are found in Hominidae, Simiiformes, Primates, Mammalia, etc.

static_heatmap.py is similar to the intheat step in taxovis. As the name says, it is not interactive, but can be used as a possible starting image for publication.

sample_taxids.py is a helper library that reads in list of NCBI taxonomic ids and a sample size and returns randomly sampled taxids. This can be used to randomly sample organisms for the taxovis heatmap. Make sure to add your email address first!