QIIME 2™ (pronounced “chime two” 🔔) is a microbiome multi-omics bioinformatics and data science platform that is trusted, free, open source, extensible, and community developed and supported.
Take a look through the thousands of published research articles, hundreds of patent applications, and tens of clinical trial records that reference the project.
Get started¶
If you have QIIME 2 Results that you’d like to interact with, you can do that without installing anything using QIIME 2 View. If you want to install and start using QIIME 2, refer to the Installing QIIME 2 section of the user documentation.
🙋 Get help, connect, and learn on the QIIME 2 Forum¶
The QIIME 2 Forum is your main resource for using QIIME 2. Running since 2016, with over 7000 registered users and over 500,000 page views per month, it’s the hub of the microbiome data science community. Find announcements of releases and workshops, discover or share microbiome-related jobs, and get technical support - all for free on the Forum. See the latest announcements below 👇.
🧐 Learn to use QIIME 2¶
The QIIME 2 user documentation is where you can find tutorials and reference content. There is a lot of excellent content here, and it is currently being refactored in JupyterBook and with Diátaxis to improve the organization and update the content to highlight the newest functionality. Check out the refactored documentation in Using QIIME 2.
🎛️ QIIME 2 makes your data analysis as efficient and accessible as possible, for you¶
If you’re a data scientist or want to integrate QIIME 2 in your existing code, checkout the Python 3 or R API.
If you prefer to work with command line tools, you can start with the command line interface.
If you prefer a graphical interface, try using QIIME 2 through Galaxy (for example, at https://
📈 Interact with QIIME 2 results and explore how they were generated¶
The new QIIME 2 View is completed re-written, enabling us to bring new features to you quicker.
For now, check out the new dynamic Visualization
gallery and consider sharing some of your favorite QIIME 2 results with the community.
📓 Perform fully reproducible bioinformatics¶
Your analysis is complex - QIIME 2 records the steps you took to be sure that your work will be reproducible by you or others.
View the provenance of your (or anyone else’s) QIIME 2 result using QIIME 2 View or Provenance Replay by running:
$ qiime tools replay-provenance \
--in-fp l6-ancom-subject.qzv \
--out-fp my-datas-provenance.bash
See the output here
#!/usr/bin/env bash
###################
# Auto-generated by qiime2 v.2024.5.0.dev0 at 05:33:08 PM on 05 May, 2024
# This document has been truncated for presentation on the qiime2.org homepage.
###################
qiime tools import \
--type 'FeatureData[Taxonomy]' \
--input-path <your data here> \
--output-path feature-data-taxonomy-0.qza
qiime tools import \
--type 'FeatureData[Sequence]' \
--input-path <your data here> \
--output-path feature-data-sequence-0.qza
qiime tools import \
--type 'EMPSingleEndSequences' \
--input-path <your data here> \
--output-path emp-single-end-sequences-0.qza
qiime feature-classifier extract-reads \
--i-sequences feature-data-sequence-0.qza \
--p-f-primer GTGCCAGCMGCCGCGGTAA \
--p-r-primer GGACTACHVGGGTWTCTAAT \
--p-trim-right 0 \
--p-trunc-len 0 \
--p-trim-left 0 \
--p-identity 0.8 \
--p-min-length 50 \
--p-max-length 0 \
--p-n-jobs 1 \
--p-batch-size auto \
--p-read-orientation both \
--o-reads reads-0.qza
# Replay attempts to represent metadata inputs accurately, but metadata .tsv
# files are merged automatically by some interfaces, rendering distinctions
# between file inputs invisible in provenance. We output the recorded
# metadata to disk to enable visual inspection.
# The following command may have received additional metadata .tsv files. To
# confirm you have covered your metadata needs adequately, review the
# original metadata, saved at './recorded_metadata/demux_emp_single_0/'
qiime demux emp-single \
--i-seqs emp-single-end-sequences-0.qza \
--m-barcodes-file <your metadata filepath>.tsv \
--m-barcodes-column <column name> \
--p-golay-error-correction \
--p-no-rev-comp-barcodes \
--p-no-rev-comp-mapping-barcodes \
--p-no-ignore-description-mismatch \
--o-per-sample-sequences per-sample-sequences-0.qza \
--o-error-correction-details XX_error_correction_details
qiime rescript dereplicate \
--i-sequences reads-0.qza \
--i-taxa feature-data-taxonomy-0.qza \
--p-mode uniq \
--p-perc-identity 1.0 \
--p-threads 1 \
--p-rank-handles greengenes \
--p-no-derep-prefix \
--o-dereplicated-taxa dereplicated-taxa-0.qza \
--o-dereplicated-sequences dereplicated-sequences-0.qza
qiime dada2 denoise-single \
--i-demultiplexed-seqs per-sample-sequences-0.qza \
--p-trunc-len 120 \
--p-trim-left 0 \
--p-max-ee 2.0 \
--p-trunc-q 2 \
--p-pooling-method independent \
--p-chimera-method consensus \
--p-min-fold-parent-over-abundance 1.0 \
--p-n-threads 1 \
--p-n-reads-learn 1000000 \
--p-hashed-feature-ids \
--o-table table-0.qza \
--o-representative-sequences representative-sequences-0.qza \
--o-denoising-stats XX_denoising_stats
qiime rescript evaluate-fit-classifier \
--i-sequences dereplicated-sequences-0.qza \
--i-taxonomy dereplicated-taxa-0.qza \
--p-reads-per-batch auto \
--p-n-jobs 1 \
--p-confidence 0.7 \
--o-classifier classifier-0.qza \
--o-evaluation XX_evaluation \
--o-observed-taxonomy XX_observed_taxonomy
# The following command may have received additional metadata .tsv files. To
# confirm you have covered your metadata needs adequately, review the
# original metadata, saved at
# './recorded_metadata/feature_table_filter_samples_0/'
qiime feature-table filter-samples \
--i-table table-0.qza \
--p-min-frequency 0 \
--p-min-features 0 \
--m-metadata-file <your metadata filepath>.tsv \
--p-where '[body-site]='"'"'gut'"'"'' \
--p-no-exclude-ids \
--p-filter-empty-features \
--o-filtered-table filtered-table-0.qza
qiime feature-classifier classify-sklearn \
--i-reads representative-sequences-0.qza \
--i-classifier classifier-0.qza \
--p-reads-per-batch auto \
--p-n-jobs 1 \
--p-pre-dispatch '2*n_jobs' \
--p-confidence 0.7 \
--p-read-orientation auto \
--o-classification classification-0.qza
qiime taxa collapse \
--i-table filtered-table-0.qza \
--i-taxonomy classification-0.qza \
--p-level 6 \
--o-collapsed-table collapsed-table-0.qza
qiime composition add-pseudocount \
--i-table collapsed-table-0.qza \
--p-pseudocount 1 \
--o-composition-table composition-table-0.qza
# The following command may have received additional metadata .tsv files. To
# confirm you have covered your metadata needs adequately, review the
# original metadata, saved at './recorded_metadata/composition_ancom_0/'
qiime composition ancom \
--i-table composition-table-0.qza \
--m-metadata-file <your metadata filepath>.tsv \
--m-metadata-column <column name> \
--p-transform-function clr \
--o-visualization visualization-0.qzv
###############################################################################
# The following QIIME 2 Results were parsed to produce this script:
# e347166b-8741-46bc-b6de-f1021b66d3b6
###############################################################################
#!/usr/bin/env python
# -----------------------------------------------------------------------------
# Auto-generated by qiime2 v.2024.5.0.dev0 at 05:54:52 PM on 05 May, 2024
# This document is a representation of the scholarly work of the creator of the
# QIIME 2 Results provided as input to this software, and may be protected by
# intellectual property law. Please respect all copyright restrictions and
# licenses governing the use, modification, and redistribution of this work.
# For User Support, post to the QIIME2 Forum at https://forum.qiime2.org.
# Instructions for use:
# 1. Open this script in a text editor or IDE. Support for Python
# syntax highlighting is helpful.
# 2. Search or scan visually for '<' or '>' characters to find places where
# user input (e.g. a filepath or column name) is required. If syntax
# highlighting is enabled, '<' and '>' will appear as syntax errors.
# 3. Search for 'FIXME' comments in the script, and respond as directed.
# 4. Remove all 'FIXME' comments from the script completely. Failure to do so
# may result in 'Missing Option' errors
# 5. Adjust the arguments to the commands below to suit your data and metadata.
# If your data is not identical to that in the replayed analysis,
# changes may be required. (e.g. sample ids or rarefaction depth)
# 6. Optional: search for 'SAVE' comments in the script, commenting out the
# `some_result.save` lines for any Results you do not want saved to disk.
# 7. Activate your replay conda environment, and confirm you have installed all
# plugins used by the script.
# 8. Run this script with `python <path to this script>`, or paste commands
# into a python interpreter or jupyter notebook for an interactive analysis
# -----------------------------------------------------------------------------
from qiime2 import Artifact
from qiime2 import Metadata
import qiime2.plugins.composition.actions as composition_actions
import qiime2.plugins.dada2.actions as dada2_actions
import qiime2.plugins.demux.actions as demux_actions
import qiime2.plugins.feature_classifier.actions as feature_classifier_actions
import qiime2.plugins.feature_table.actions as feature_table_actions
import qiime2.plugins.rescript.actions as rescript_actions
import qiime2.plugins.taxa.actions as taxa_actions
feature_data_taxonomy_0 = Artifact.import_data(
'FeatureData[Taxonomy]',
<your data here>,
)
# SAVE: comment out the following with '# ' to skip saving this Result to disk
feature_data_taxonomy_0.save('feature_data_taxonomy_0')
feature_data_sequence_0 = Artifact.import_data(
'FeatureData[Sequence]',
<your data here>,
)
# SAVE: comment out the following with '# ' to skip saving this Result to disk
feature_data_sequence_0.save('feature_data_sequence_0')
emp_single_end_sequences_0 = Artifact.import_data(
'EMPSingleEndSequences',
<your data here>,
)
# SAVE: comment out the following with '# ' to skip saving this Result to disk
emp_single_end_sequences_0.save('emp_single_end_sequences_0')
reads_0, = feature_classifier_actions.extract_reads(
sequences=feature_data_sequence_0,
f_primer='GTGCCAGCMGCCGCGGTAA',
r_primer='GGACTACHVGGGTWTCTAAT',
trim_right=0,
trunc_len=0,
trim_left=0,
identity=0.8,
min_length=50,
max_length=0,
n_jobs=1,
batch_size='auto',
read_orientation='both',
)
# SAVE: comment out the following with '# ' to skip saving Results to disk
reads_0.save('reads_0')
# Replay attempts to represent metadata inputs accurately, but metadata .tsv
# files are merged automatically by some interfaces, rendering distinctions
# between file inputs invisible in provenance. We output the recorded metadata
# to disk to enable visual inspection.
# The following command may have received additional metadata .tsv files. To
# confirm you have covered your metadata needs adequately, review the original
# metadata, saved at './recorded_metadata/demux_emp_single_0/'
# NOTE: You may substitute already-loaded Metadata for the following, or cast a
# pandas.DataFrame to Metadata as needed.
barcodes_0_md = Metadata.load(<your metadata filepath>)
barcodes_0_mdc_0_mdc = barcodes_0_md.get_column('<column name>')
per_sample_sequences_0, _ = demux_actions.emp_single(
seqs=emp_single_end_sequences_0,
barcodes=barcodes_0_mdc_0_mdc,
golay_error_correction=True,
rev_comp_barcodes=False,
rev_comp_mapping_barcodes=False,
ignore_description_mismatch=False,
)
# SAVE: comment out the following with '# ' to skip saving Results to disk
per_sample_sequences_0.save('per_sample_sequences_0')
dereplicated_sequences_0, dereplicated_taxa_0 = rescript_actions.dereplicate(
sequences=reads_0,
taxa=feature_data_taxonomy_0,
mode='uniq',
perc_identity=1.0,
threads=1,
rank_handles='greengenes',
derep_prefix=False,
)
# SAVE: comment out the following with '# ' to skip saving Results to disk
dereplicated_taxa_0.save('dereplicated_taxa_0')
dereplicated_sequences_0.save('dereplicated_sequences_0')
table_0, representative_sequences_0, _ = dada2_actions.denoise_single(
demultiplexed_seqs=per_sample_sequences_0,
trunc_len=120,
trim_left=0,
max_ee=2.0,
trunc_q=2,
pooling_method='independent',
chimera_method='consensus',
min_fold_parent_over_abundance=1.0,
n_threads=1,
n_reads_learn=1000000,
hashed_feature_ids=True,
)
# SAVE: comment out the following with '# ' to skip saving Results to disk
table_0.save('table_0')
representative_sequences_0.save('representative_sequences_0')
classifier_0, _, _ = rescript_actions.evaluate_fit_classifier(
sequences=dereplicated_sequences_0,
taxonomy=dereplicated_taxa_0,
reads_per_batch='auto',
n_jobs=1,
confidence=0.7,
)
# SAVE: comment out the following with '# ' to skip saving Results to disk
classifier_0.save('classifier_0')
# The following command may have received additional metadata .tsv files. To
# confirm you have covered your metadata needs adequately, review the original
# metadata, saved at './recorded_metadata/feature_table_filter_samples_0/'
# NOTE: You may substitute already-loaded Metadata for the following, or cast a
# pandas.DataFrame to Metadata as needed.
metadata_0_md = Metadata.load(<your metadata filepath>)
filtered_table_0, = feature_table_actions.filter_samples(
table=table_0,
min_frequency=0,
min_features=0,
metadata=metadata_0_md,
where="[body-site]='gut'",
exclude_ids=False,
filter_empty_features=True,
)
# SAVE: comment out the following with '# ' to skip saving Results to disk
filtered_table_0.save('filtered_table_0')
classification_0, = feature_classifier_actions.classify_sklearn(
reads=representative_sequences_0,
classifier=classifier_0,
reads_per_batch='auto',
n_jobs=1,
pre_dispatch='2*n_jobs',
confidence=0.7,
read_orientation='auto',
)
# SAVE: comment out the following with '# ' to skip saving Results to disk
classification_0.save('classification_0')
collapsed_table_0, = taxa_actions.collapse(
table=filtered_table_0,
taxonomy=classification_0,
level=6,
)
# SAVE: comment out the following with '# ' to skip saving Results to disk
collapsed_table_0.save('collapsed_table_0')
composition_table_0, = composition_actions.add_pseudocount(
table=collapsed_table_0,
pseudocount=1,
)
# SAVE: comment out the following with '# ' to skip saving Results to disk
composition_table_0.save('composition_table_0')
# The following command may have received additional metadata .tsv files. To
# confirm you have covered your metadata needs adequately, review the original
# metadata, saved at './recorded_metadata/composition_ancom_0/'
# NOTE: You may substitute already-loaded Metadata for the following, or cast a
# pandas.DataFrame to Metadata as needed.
metadata_1_md = Metadata.load(<your metadata filepath>)
metadata_1_mdc_0_mdc = metadata_1_md.get_column('<column name>')
visualization_0_viz, = composition_actions.ancom(
table=composition_table_0,
metadata=metadata_1_mdc_0_mdc,
transform_function='clr',
)
# SAVE: comment out the following with '# ' to skip saving Results to disk
visualization_0_viz.save('visualization_0_viz')
# -----------------------------------------------------------------------------
# The following QIIME 2 Results were parsed to produce this script:
# e347166b-8741-46bc-b6de-f1021b66d3b6
# -----------------------------------------------------------------------------
🔌 Build, support, and publish your own QIIME 2 plugins¶
Developing with QIIME 2 is your main resource for learning QIIME 2 plugin development, starting with a 7 lesson tutorial where you’ll build your first QIIME 2 plugin from scratch. New content is added regularly.
📚 Discover tools built by others!¶
Coming soon: We are currently refactoring the QIIME 2 Library to make it more useful for developers and users. More on this throughout 2024!
🙏 Funding¶
QIIME 2 is funded in part by NIH National Cancer Institute Informatics Technology for Cancer Research grant 1U24CA248454-01, and by grant DAF2019-207342 from the Chan Zuckerberg Initiative (CZI) DAF, an advised fund of Silicon Valley Community Foundation (CZI grant DOI: 10.37921/862772dbrrej; funder DOI 10.13039/100014989).
This website was built with MyST Markdown, which is supported in part with funding from the Alfred P. Sloan Foundation.
Initial support for the development of QIIME 2 was provided through a grant from the National Science Foundation.