q2-fondue

q2-fondue is a QIIME 2 plugin for programmatic access to sequences and metadata from NCBI Sequence Read Archive (SRA). It enables user-friendly acquisition, re-use, and management of public nucleotide sequence (meta)data while adhering to open data principles.

Installation¶

There are multiple options to install q2-fondue (v2024.5 or higher) - each targeted towards different needs:

(To install q2-fondue with a version <= 2023.7 see this section.)

Option 1: Install q2-fondue with QIIME 2 moshpit distribution¶

q2-fondue is a part of the QIIME 2 moshpit distribution and you can install it as outlined in the QIIME 2 installation instructions. After that, don’t forget to run the mandatory configuration step!

Option 2: Install q2-fondue within a QIIME 2 amplicon conda environment¶

• Install the QIIME 2 amplicon distribution within a conda environment as described in the official QIIME 2 documentation.
• Activate the QIIME 2 environment (v2024.5 or higher) and install q2-fondue within while making sure that the used conda channel matches the version of the QIIME 2 environment (replace below {ENV_VERSION} with the version number of your QIIME 2 environment):

conda activate qiime2-amplicon-{ENV_VERSION}
mamba install -y \
   -c https://packages.qiime2.org/qiime2/{ENV_VERSION}/metagenome/released/ \
   -c conda-forge -c bioconda -c defaults \
   q2-fondue

• Now, don’t forget to run the mandatory configuration step!

Option 3: Minimal fondue environment¶

• Start with installing mamba in your base environment:

conda install mamba -n base -c conda-forge

• Create and activate a conda environment with the required dependencies:

mamba create -y -n fondue \
   -c https://packages.qiime2.org/qiime2/2024.5/metagenome/released/ \
   -c conda-forge -c bioconda -c defaults \
   q2cli q2-fondue

conda activate fondue

• Now, don’t forget to run the mandatory configuration step!

Note: You can replace the version number 2024.5 with later releases if they are already available.

Mandatory configuration for all three options¶

• Refresh the QIIME 2 CLI cache and see that everything worked:

qiime dev refresh-cache
qiime fondue --help

• Run the vdb-config tool to make sure the wrapped SRA Toolkit is configured on your system. The command below will open the configuration interface - everything should be already configured, so you can directly exit by pressing x (this step is still required to ensure everything is working as expected). Feel free to adjust the configuration, if you need to change e.g. the cache location. For more information see here.

vdb-config -i

• In case you need to configure a proxy server, run the following command (this can also be done using the graphical interface described above):

vdb-config --proxy <your proxy URL> --proxy-disable no

Installing q2-fondue with older versions¶

To install fondue with a version <= 2023.7 in a minimal environment run the following command inserting the respective version number {ENV_VERSION}:

mamba create -y -n fondue \
   -c https://packages.qiime2.org/qiime2/{ENV_VERSION}/tested/ \
   -c conda-forge -c bioconda -c defaults \
   q2cli q2-fondue

conda activate fondue

Now, don’t forget to run the mandatory configuration step!

Alternatively, a minimal Docker image is available to run q2-fondue==v2023.7:

• Install Docker with the linked instructions
• Pull the q2-fondue Docker image:

docker pull linathekim/q2-fondue:2023.7

• Within the container, refresh the QIIME 2 CLI cache to see that everything worked:

qiime dev refresh-cache
qiime fondue --help

• If you need to configure a proxy server, run the following command:

vdb-config --proxy <your proxy URL> --proxy-disable no

Space requirements¶

Running q2-fondue requires space in the temporary (TMPDIR) and output directory. The space requirements for the output directory can be estimated by inserting the run or project IDs in the SRA Run Selector. To estimate the space requirements for the temporary directory, multiply the output directory space requirement by a factor of 10. The current implementation of q2-fondue requires you to have a minimum of 2 GB of available space in your temporary directory.

To find out which temporary directory is used by Qiime 2, you can run echo $TMPDIR in your terminal. If this command returns an empty string, the assigned temporary directory equals the OS’s default temporary directory (usually /tmp) . To re-assign your temporary directory to a location of choice, run export TMPDIR=Location/of/choice.

Usage¶

Available actions¶

q2-fondue provides a couple of actions to fetch and manipulate nucleotide sequencing data and related metadata from SRA as well as an action to scrape run, study, BioProject, experiment and sample IDs from a Zotero web library. Below you will find a list of available actions and their short descriptions.

[*]: Supported IDs include run, study, BioProject, experiment and study IDs.

The next sections give a brief introduction to the most important actions in q2-fondue. More detailed instructions, background information and examples can be found in the associated tutorial.

Import accession IDs¶

All q2-fondue actions which fetch data from SRA require the list of run, study, BioProject, experiment or study IDs to be provided as a QIIME 2 artifact of NCBIAccessionIDs semantic type. You can either import an existing list of IDs (1.) or scrape a Zotero web library collection to obtain these IDs (2.).

1. To import an existing list of IDs into a NCBIAccessionIDs artifact simply run:

qiime tools import \
              --type NCBIAccessionIDs \
              --input-path ids.tsv \
              --output-path ids.qza

where:

• --input-path is a path to the TSV file containing run or project IDs.
• --output-path is the output artifact.

Note: the input TSV file needs to consist of a single column named “ID”.

2. To scrape all run, study, BioProject, experiment and sample IDs from an existing web Zotero library collection into a NCBIAccessionIDs artifact run, you can use the scrape-collection method. Before running it, you have to set three environment variables linked to your Zotero account:

• ZOTERO_TYPE is the Zotero library type ‘user’ or ‘group’.
• ZOTERO_USERID is a valid Zotero user ID. If ZOTERO_TYPE is ‘user’ it can be retrieved from section ‘your user_id for use in API calls’ in https://www.zotero.org/settings/keys. If ZOTERO_TYPE is ‘group’ it can be obtained by hovering over group name in https://www.zotero.org/groups/.
• ZOTERO_APIKEY is a valid Zotero API user key created at https://www.zotero.org/settings/keys/new (checking “Allow library access” and for ‘group’ library “Read/Write” permissions).

To set these environment variables run the following commands in your terminal for each of the three required variables: export ZOTERO_TYPE=<your library type> or create a .env file with the environment variable assignment. For the latter option, make sure to ignore this file in version control (add to .gitignore).

Note: To retrieve all required entries from Zotero, you must be logged in. Also, to allow for the scrape-collection action to work, make sure you enable file syncing on your Zotero account (see section “File Syncing” here) and only attempt to use the action once all attachments were synchronized with your Web Library.

qiime fondue scrape-collection \
              --p-collection-name collection_name \
              --o-run-ids run_ids.qza \
              --o-study-ids study_ids.qza \
              --o-bioproject-ids bioproject_ids.qza \
              --o-experiment-ids experiment_ids.qza \
              --o-sample-ids sample_ids.qza --verbose

where:

• --p-collection-name is the name of the collection to be scraped.
• --o-run-ids is the output artifact containing the scraped run IDs.
• --o-study-ids is the output artifact containing the scraped study IDs.
• --o-bioproject-ids is the output artifact containing the scraped BioProject IDs.
• --o-experiment-ids is the output artifact containing the scraped experiment IDs.
• --o-sample-ids is the output artifact containing the scraped sample IDs.

3. To retrieve run accession IDs based on a text search query (performed on the BioSample database) you can use the get-ids-from-query method:

qiime fondue get-ids-from-query \
              --p-query "txid410656[Organism] AND \"public\"[Filter] AND (chicken OR poultry)" \
              --p-email [email protected] \
              --p-threads 2 \
              --o-ids run_ids.qza \
              --verbose

where:

• --p-query is the text search query to be executed on the BioSample database.
• --p-email is your email address (required by NCBI).
• --p-threads is the number of parallel download jobs (defaults to 1).
• --o-ids is the output artifact containing the retrieved run IDs.

Fetching metadata¶

To fetch metadata associated with a set of output IDs, execute the following command:

qiime fondue get-metadata \
              --i-accession-ids ids.qza \
              --p-threads 1 \
              --p-email [email protected] \
              --o-metadata output_metadata.qza \
              --o-failed-runs failed_IDs.qza

where:

• --i-accession-ids is an artifact containing run, study, BioProject, experiment or sample IDs
• --p-threads is a number of parallel download jobs (defaults to 1)
• --p-email is your email address (required by NCBI)
• --o-metadata is the output metadata artifact
• --o-failed-runs is the list of all run IDs for which fetching metadata failed, with their corresponding error messages

The resulting artifact --o-metadata will contain a TSV file with all the available metadata fields for all of the requested runs. If metadata for some run IDs failed to download they are returned in the --o-failed-runs artifact, which can be directly inputted as --i-accession-ids to a subsequent get-metadata command. To pass associated DOI names for the failed runs, provide the table of accession IDs with associated DOI names as --o-linked-doi to the get-metadata command.

Fetching sequences¶

To get openly accessible single- and paired-end sequences associated with a number of IDs, execute this command:

qiime fondue get-sequences \
              --i-accession-ids ids.qza \
              --p-email [email protected] \
              --o-single-reads output_dir_single \
              --o-paired-reads output_dir_paired \
              --o-failed-runs output_failed_ids

where:

• --i-accession-ids is an artifact containing run, study, BioProject, experiment or sample IDs
• --p-email is your email address (required by NCBI)
• --o-single-reads is the output artifact containing single-read sequences
• --o-paired-reads is the output artifact containing paired-end sequences
• --o-failed-runs is the output artifact containing run IDs that failed to download

The resulting sequence artifacts (--o-single-reads and --o-paired-reads) will contain the fastq.gz files of the sequences, metadata.yml and MANIFEST files. If one of the provided IDs only contains sequences of one type (e.g. single-read sequences) then the other artifact (e.g. artifact with paired-end sequences) contains empty sequence files with dummy ID starting with xxx_. Similarly, if none of the requested sequences failed to download, the corresponding artifact will be empty.

If some run IDs failed to download they are returned in the --o-failed-runs artifact, which can be directly inputted as --i-accession-ids to a subsequent get-sequence command.

Special case: Fetching restricted access sequences with a dbGAP repository key¶

To get access to the respective dbGaP repository key users must first apply for approval and then retrieve the key from dbGAP (see prerequisites described here).

To retrieve sequencing data using the acquired dbGAP repository key, without revealing the sensitive key, set the filepath to the stored key as an environment variable. You can either do this by running the following command in your terminal export KEY_FILEPATH=<path to key> or by adding the variable assignment to your .env file. For the latter option, make sure to ignore this file in version control (add to .gitignore). Having set the filepath of the key as an environment variable you can fetch the sequencing data by running get-sequences with the parameter --p-restricted-access:

qiime fondue get-sequences \
              --i-accession-ids ids.qza \
              --p-email [email protected] \
              --p-restricted-access \
              --output-dir output_path

Note: Fetching metadata with a dbGAP repository key is not supported. Hence, this flag is only available in the get-sequences action (and not in the get-metadata and get-all actions).

Fetching metadata and sequences¶

To fetch both sequence-associated metadata and sequences associated with the provided IDs, execute this command:

qiime fondue get-all \
              --i-accession-ids ids.qza \
              --p-email [email protected] \
              --output-dir output-dir-name

where:

• --i-accession-ids is an artifact containing run, study, BioProject, experiment or sample IDs
• --p-email is your email address (required by NCBI)
• --output-dir directory where the downloaded metadata, sequences and IDs for failed downloads are stored as QIIME 2 artifacts

Parallelization¶

q2-fondue makes use of the powerful parsl-based parallelization provided by the Q2 framework. By default, if no additional options are provided, all the data will be fetched sequentially, run by run. To speed up the process, you can configure the actions to use multiple workers with each worker processing one run. Here is an example of the parsl config which will allow you to spin up 4 workers on your local machine:

[parsl]

[[parsl.executors]]
class = "HighThroughputExecutor"
label = "default"
max_workers = 1

[parsl.executors.provider]
class = "LocalProvider"
max_blocks = 4

You can also find this sample config file here. You can pass this config file to the get-sequences and get-all actions using the --parallel-config flag. For more information on parsl-based parallelization in QIIME 2, please refer to the documentation.

Note: Do not request more than 1 CPU per parsl worker - each worker will only process one run at a time so requesting more than 1 CPU will not speed up the process.

Downstream analysis in QIIME 2¶

For more information on how to use q2-fondue outputs within the QIIME 2 ecosystem see section Downstream analysis in QIIME 2 in the tutorial.

Exporting data for downstream analyses outside of QIIME 2¶

Some downstream analyses may need to rely on tools outside of QIIME 2. Since q2-fondue outputs can be transformed directly into FASTQ and other interoperable formats, there are no restrictions for users when using these tools. Note that the exported files will no longer contain integrated provenance information (which is unique to QIIME 2 Artifacts), but this metadata can be exported also and the original artifacts will retain the provenance data for traceability purposes.

To learn more on how to prepare q2-fondue outputs for further analysis outside of QIIME 2 see tutorial section Prepare downstream analysis outside of QIIME 2.

Getting Help¶

Problem? Suggestion? Technical errors and user support requests can be filed on the QIIME 2 Forum.

Citation¶

If you use fondue in your research, please cite the following:

Michal Ziemski, Anja Adamov, Lina Kim, Lena Flörl, Nicholas A. Bokulich. 2022. Reproducible acquisition, management, and meta-analysis of nucleotide sequence (meta)data using q2-fondue. _Bioinformatics; doi: https://doi.org/10.1093/bioinformatics/btac639

License¶

q2-fondue is released under a BSD-3-Clause license. See LICENSE for more details.