enable LINs for remaining tax utilities
#2499
Comments
8 tasks
|
tax genome is in #2519 |
bluegenes
added a commit
that referenced
this issue
Apr 5, 2023
) ## Add taxonomic utilities for LINs; enable and test `tax metagenome` With taxonomy refactoring (#2437, #2439, #2443, #2446, #2466, #2467), we are (mostly) no longer tied to named ranks. Here, I add a class for LIN taxonomies and use it within `tax metagenome` to allow summarization up LINs and reporting at specified `lingroups`. With this PR, users can now use the flag `--lins` to read and use `lin` taxonomies from the provided tax (`-t`, `--taxonomy`) file. If used, `sourmash tax` will look for a `lin` column in the taxonomy file instead of looking for `superkingdom`...`strain` columns. The `lin` column should contain `;`-separated LINs, preferably with a standard number of positions (e.g. all 20 positions in length or all 10 positions in length). For `tax metagenome`: By default, `tax metagenome` will summarize up _all_ available ranks/LIN positions. If a `lingroup` file is provided, we will also report a subset of this summary: just the LIN prefixes that match groups in the `lingroup` file. The `lingroup` file requires two columns in any order: `name`, the name of the group, and `lin`, the lin prefix of the group. The prefix will be used to select results from the full summary for reporting. The `lingroup` format will build a file with the following name: `{base}.lingroup.tsv`, where `{base}` is the name provided via the `-o`,` --output-base` option. ## Demo / Tutorial A draft tutorial is available [here](https://sourmash--2469.org.readthedocs.build/en/2469/tutorial-lin-taxonomy.html). Note that it does not contain the installation info for this branch (see below). You can run the interactive version via binder [here](https://mybinder.org/v2/gh/bluegenes/2023-demo-sourmash-LIN/HEAD?labpath=sourmash-lin-demo.ipynb) ## Testing ### Option A: Use the Demo Binder You can test via the [binder](https://mybinder.org/v2/gh/bluegenes/2023-demo-sourmash-LIN/HEAD?labpath=sourmash-lin-demo.ipynb). You can add new cells or modify any existing cells, and even download additional files for testing. The downside is that you'll have to make sure to download and save your results, since the binder won't save them for you. ### Option B: Alternatively, install on your own computer/cluster: Here is one way to test this code before it gets fully integrated into sourmash: - If you don't have conda, I'd recommend installing `mamba`, [instructions here](https://mamba.readthedocs.io/en/latest/installation.html) instead. - if you do have `mamba`, replace the word `conda` with `mamba` in the following commands. Download an environment file that points to this branch: ``` curl -JLO https://raw.githubusercontent.com/bluegenes/2023-demo-sourmash-LIN/main/sourmashLIN.yml ``` Create a virtual environment using this file: ``` conda env create -f sourmashLIN.yml ``` Activate that environment: ``` conda activate smashLIN ``` make sure `--lins` is in the `--help` for `sourmash tax metagenome`: ``` sourmash tax metagenome --help ``` ## Command to run The command to run is this one: ``` sourmash tax metagenome -g $gather_csv -t $taxonomy_csv \ --lins --lingroup $lingroups_csv ``` ## Types of files you'll need 1. sketches of query metagenome 2. sketches of reference genomes (database) 3. taxonomy file with LIN information (two columns required: `ident`, `lin`) 4. lingroup information file (two columns required: `name`, `lin`) To exit the environment when you're done testing, use `conda deactivate` > Reminder, if you have `mamba`, you can use it in place of `conda` in the commands above. example `lingroup` output format. Note that the `1;0`.. paths are always grouped together, but may come before or after the `0;0` and `2;0` groups. ``` name lin percent_containment num_bp_contained lg3 2;0;0 1.56 192000 lg1 0;0;0 5.82 714000 lg2 1;0;0 5.05 620000 lg3 1;0;1 0.65 80000 lg4 1;0;1;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0 0.65 80000 ``` ``` name lin percent_containment num_bp_contained lg2 1;0;0 5.05 620000 lg3 1;0;1 0.65 80000 lg4 1;0;1;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0;0 0.65 80000 lg1 0;0;0 5.82 714000 lg3 2;0;0 1.56 192000 ``` ## A few implementation details: - In `tax_utils.py`, I add a `LINLineageInfo` class for using and manipulated LIN taxonomies. It implements new methods to enable specifically reading in `LIN` taxonomies into the class, but otherwise uses the taxonomic utilities available in `BaseLineageInfo`, e.g. taxonomic summarization up ranks, assessing whether two taxonomies are a match at a given rank. - In `tax_utils.py`, I add functionality for reading `lingroup` information and reporting taxonomic summarization specifically at these ranks. Changes and Additions: - [x] Add `LINLineageInfo` for working with `LIN` taxonomies - [x] Add method for reading `LIN`s into `LineageDB` - [x] Add methods for reading `LINgroups` and summarizing to these - [x] Add LineageTree that can use `LineageInfo` to perform `build_tree`, `find_lca` functions (originally in `lca_utils.py`) and produce an ordered list of lineage paths - [x] Add code + tests to use `LIN`s taxonomy in: - [x] tax metagenome - [x] tax annotate - [x] tax summarize The following require additional changes and will be punted to an issue/separate PR (see #2499): - tax genome - tax prepare - tax grep --------- Co-authored-by: C. Titus Brown <titus@idyll.org>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
with #2469 and #2496 we will introduce utilities for using the
LINtaxonomic framework and enable it for severaltaxutilities, most critically,tax metagenome.Three remain to fully utilize
LINs:The text was updated successfully, but these errors were encountered: