Metagenomics with QIIME 2 - Spring 2024 Leiden & Zurich Tutorial

Metagenomics with QIIME 2 - Spring 2024 Leiden & Zurich Tutorial#

Warning

Metagenomics analysis with QIIME 2 is in alpha release. This means that results you generate should be considered preliminary, and NOT PUBLICATION QUALITY. Additionally, interfaces are subject to change, and those changes may be backward incompatible (meaning that a command or file that works in one version of the QIIME 2 Shotgun Metagenomics distribution may not work in the next version of that distribution).

Setup#

Before we dive into the tutorial, let’s talk about our workshop server directory structure.

Directory structure#

<your home directory>
└ workshop
   ├ reads
   ├ contigs
   └ sample-metadata.tsv

Let’s verify that QIIME 2 is working by calling qiime.

qiime info

Before we start our analyses, we want to create sub-directories for each type of data we’re examining to keep things organized. Within the workshop directory, we’ll create the following sub-directories:

mkdir reads

mkdir contigs

Required databases

In order to perform the taxonomic and functional annotation, we will need a couple of different reference databases. Below you will find instructions on how to download these databases using respective QIIME 2 actions.

Kraken 2/Bracken database

qiime moshpit build-kraken-db \
    --p-collection standard \
    --o-kraken2-database ./moshpit_tutorial/cache:kraken_standard \
    --o-bracken-database ./moshpit_tutorial/cache:bracken_standard \
    --verbose

EggNOG databases

qiime moshpit fetch-diamond-db \
    --o-diamond-db ./moshpit_tutorial/cache:eggnog_diamond_full \
    --verbose

qiime moshpit fetch-eggnog-db \
    --o-eggnog-db ./moshpit_tutorial/cache:eggnog_annot_full \
    --verbose

Data retrieval from SRA

The data we are using in this tutorial can be fetched from the SRA repository using the q2-fondue plugin. We need to import the accession IDs into a QIIME 2 artifact:

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

We can then use the get-sequences action to download the data (please insert your e-mail address in the --p-email parameter):

qiime fondue get-sequences \
    --i-accession-ids ./moshpit_tutorial/ids.qza \
    --p-n-jobs 16 \
    --p-email you@tutorial.com \
    --o-single-reads ./moshpit_tutorial/cache:reads_single \
    --o-single-paired ./moshpit_tutorial/cache:reads_paired \
    --o-failed-runs ./moshpit_tutorial/cache:failed_runse \
    --verbose

This will download all the sequences into the QIIME 2 cache. It is a lot of data, so keep in mind that depending on your network speed, this might take a while.

Metadata#

First, let’s grab our sample metadata!

wget -O sample-metadata.tsv https://polybox.ethz.ch/index.php/s/79s2cQry8Ll0FGq/download

Next, let’s take a look at the metadata by visualizing it in QIIME 2.

qiime metadata tabulate \
  --m-input-file sample-metadata.tsv \
  --o-visualization sample-metadata.qzv

Note

We will be using the newest version of QIIME 2 View (our Results viewer) to examine our QIIME 2 Visualizations. The URL for this version is as follows:

https://q2view-svelte.pages.dev/

While this version is still in alpha release, we wanted to use this opportunity to share it with a live audience, and get any feedback from you on usability. Please let us know if you run into any issues while using it! If you find yourself blocked on any of the tutorial steps, please revert to using our current release version of QIIME 2 View (view.qiime2.org).

In order to proceed, we first need to grab the URL for each visualization from the workshop server. For each visualization, you’ll navigate to:

https://workshop-server.qiime2.org/[your-username]/

From here, you’ll right click on the hyperlink associated with the visualization you’d like to view, copy the URL, navigate to QIIME 2 View and paste this URL under a file from the web. Note that you can also download any of these files directly from the server by left clicking on the hyperlink associated with the visualization you’d like to download.

Read-based analysis#

qiime moshpit classify-kraken2 \
	--i-seqs ./moshpit_tutorial/cache:workshop-reads \
	--i-kraken2-db ./moshpit_tutorial/cache:kracken_standard \
	--p-threads 40 \
	--p-confidence 0.6 \
	--p-minimum-base-quality 20 \
	--o-hits ./moshpit_tutorial/cache:workshop_kraken_db_hits \
	--o-reports ./moshpit_tutorial/cache:workshop_kraken_db_reports \
	--p-report-minimizer-data \
	--use-cache ./moshpit_tutorial/cache \
	--parallel-config slurm_config.toml \
    	--verbose \
    	--p-memory-mapping False ##set to False to shorten runtime

qiime moshpit estimate-bracken \
    --i-bracken-db ./moshpit_tutorial/cache:bracken_standard \
    --p-read-len 100 \
    --i-kraken-reports ./moshpit_tutorial/cache:workshop_kraken_db_reports \
    --o-reports ./moshpit_tutorial/kraken-outputs/bracken-reports.qza \
    --o-taxonomy ./moshpit_tutorial/kraken-outputs/taxonomy-bracken.qza \
    --o-table ~./moshpit_tutorial/kraken-outputs/table-bracken.qza

Obtaining your Feature Table and Taxonomy Table#

We are going to look at taxonomic annotations for our read-based analysis. In order to do that, we need to download our read table and taxonomy that we generated using Bracken.

wget -O ./reads/bracken-feature-table.qza https://polybox.ethz.ch/index.php/s/4Y1IGtZHTzo1KTi/download

wget -O ./reads/bracken-taxonomy.qza https://polybox.ethz.ch/index.php/s/haWDZzLcJsuiI9b/download

Before moving forward in our analysis, let’s take a look at a visual summary of our feature table and re-examine our metadata summary. This will help to orient us with what our original data looks like, and put subsequent downstream steps in context.

qiime feature-table summarize \
  --i-table ./reads/bracken-feature-table.qza \
  --m-sample-metadata-file sample-metadata.tsv \
  --o-visualization ./reads/bracken-feature-table.qzv

Feature Table Filtering#

First, we need to remove samples that are not part of the autoFMT study from the feature table. We’ll identify these samples using the metadata. Specifically, this step filters samples that do not contain a value in the autoFmtGroup column in the metadata.

qiime feature-table filter-samples \
  --i-table ./reads/bracken-feature-table.qza \
  --m-metadata-file sample-metadata.tsv \
  --p-where 'autoFmtGroup IS NOT NULL' \
  --o-filtered-table ./reads/bracken-autofmt-feature-table.qza

Taxa Barplot Creation#

Now we will use our filtered table and taxonomy to create a taxonomic barplot.

qiime taxa barplot \
  --i-table ./reads/bracken-autofmt-feature-table.qza \
  --i-taxonomy ./reads/bracken-taxonomy.qza \
  --m-metadata-file sample-metadata.tsv \
  --o-visualization ./reads/taxa-bar-plot-autofmt-reads.qzv

Differential Abundance#

Feature Table Preparation#

We’d now like to perform differential abundance analysis on our data, which aims to find the differences in the abundance of each taxa between two classes of subjects or samples, assigning a significance value to each comparison. For this analysis, we’ll be using ANCOM-BC (which cannot be run on repeated measures), so we’ll first need to filter to one timepoint. Here we will filter to the “peri” timepoint.

qiime feature-table filter-samples \
  --i-table ./reads/bracken-autofmt-feature-table.qza \
  --m-metadata-file sample-metadata.tsv \
  --p-where "[categorical-time-relative-to-fmt]='peri'" \
  --o-filtered-table ./reads/peri-fmt-table.qza

Now let’s visualize our filtered table!

qiime feature-table summarize \
  --i-table ./reads/peri-fmt-table.qza \
  --m-sample-metadata-file sample-metadata.tsv \
  --o-visualization ./reads/peri-fmt-table.qzv

Upon examination of our filtered table, it looks like there is still a subject that has two samples at our peri timepoint. Let’s filter that out!

echo SampleID > ./reads/samples-to-remove.tsv
echo SRR14092317 >> ./reads/samples-to-remove.tsv

qiime feature-table filter-samples \
  --i-table ./reads/peri-fmt-table.qza \
  --m-metadata-file ./reads/samples-to-remove.tsv \
  --p-exclude-ids \
  --o-filtered-table ./reads/id-filtered-peri-fmt-table.qza

qiime feature-table summarize \
  --i-table ./reads/id-filtered-peri-fmt-table.qza \
  --m-sample-metadata-file sample-metadata.tsv \
  --o-visualization ./reads/id-filtered-peri-fmt-table.qzv

Now we should collapse our table so our features are grouped at the species level.

qiime taxa collapse \
--i-table ./reads/id-filtered-peri-fmt-table.qza \
--i-taxonomy ./reads/bracken-taxonomy.qza \
--p-level 8 \
--o-collapsed-table ./reads/collapsed-8-id-filtered-peri-fmt-table.qza

Differential Abundance with ANCOM-BC#

Now let’s run ANCOM-BC! This is a great way to take a look at what features are either enriched or depleted, relative to a reference group of our choosing. Since we’d like to look at the differential abundance of the autoFmtGroup, we’ll set our reference group to control.

 qiime composition ancombc \
  --i-table ./reads/collapsed-8-id-filtered-peri-fmt-table.qza  \
  --m-metadata-file sample-metadata.tsv \
  --p-formula autoFmtGroup \
  --p-reference-levels autoFmtGroup::control \
  --o-differentials ./reads/differentials-peri-autofmt.qza

qiime composition da-barplot \
  --i-data ./reads/differentials-peri-autofmt.qza \
  --p-significance-threshold 0.05 \
  --p-level-delimiter ";" \
  --o-visualization ./reads/differentials-peri-autofmt.qzv

Contig-based analysis#

Assemble Reads into Contigs with MEGAHIT

The first step in recovering metagenome-assembled genomes (MAGs) is genome assembly itself. There are many genome assemblers available, two of which you can use through our QIIME 2 plugin - here, we will use MEGAHIT. MEGAHIT takes short DNA sequencing reads, constructs a simplified De Bruijn graph, and generates longer contiguous sequences called contigs, providing valuable genetic information for the next steps of our analysis.

qiime assembly assemble-megahit \
    --i-seqs "./moshpit_tutorial/cache:reads_no_host" \
    --p-presets "meta-sensitive" \
    --p-num-cpu-threads 64 \
    --p-num-partitions 4 \
    --o-contigs "./moshpit_tutorial/cache:contigs" \
    --verbose

Contig QC with QUAST

Once the reads are assembled into contigs, we can use QUAST to evaluate the quality of our assembly.

qiime assembly evaluate-contigs \
    --i-contigs "./moshpit_tutorial/cache:contigs" \
    --p-threads 128 \
    --p-memory-efficient \
    --o-visualization "./moshpit_tutorial/results/contigs.qzv" \
    --verbose

QUAST QC#

After assembling our reads into contigs, let’s have a look at our contig quality metrics.

wget -O ./contigs/quast-qc.qzv https://polybox.ethz.ch/index.php/s/XyZfYkDEHh1nHZq/download

Obtaining your Feature Table and Taxonomy Table#

qiime moshpit classify-kraken2 \
    --i-seqs "./moshpit_tutorial/cache:megahit-contigs" \
    --i-kraken2-db "./moshpit_tutorial/cache:kraken_standard" \
    --p-threads 48 \
    --p-confidence 0.6 \
    --p-minimum-base-quality 20 \
    --p-num-partitions 4 \
    --o-reports "./moshpit_tutorial/cache:kraken_reports_contigs" \
    --o-hits "./moshpit_tutorial/cache:kraken_hits_contigs" \
    --verbose

qiime moshpit kraken2-to-features \
  --i-reports "./moshpit_tutorial/cache:kraken_reports_contigs" \
  --o-table "./moshpit_tutorial/cache:kraken_feature_table_contigs" \
  --o-taxonomy "./moshpit_tutorial/cache:kraken_taxonomy_contigs" \
  --verbose

Similar to our read-based analysis, we are now going to look at taxonomic annotations for our contig-based analysis. In order to do that, we need to download our contig feature table and taxonomy that we generated using Kraken2.

wget -O ./contigs/kraken2-presence-absence-contigs.qza https://polybox.ethz.ch/index.php/s/OYL590hv7eZJPWS/download

wget -O ./contigs/kraken2-taxonomy-contigs.qza https://polybox.ethz.ch/index.php/s/Wk0nsgQfEjgdabc/download

Feature Table Filtering#

Now that we have acquired our contig feature table, we will also remove samples that are not part of the autoFMT study (following the same approach we used in our read-based analysis).

qiime feature-table filter-samples \
  --i-table ./contigs/kraken2-presence-absence-contigs.qza \
  --m-metadata-file sample-metadata.tsv \
  --p-where 'autoFmtGroup IS NOT NULL' \
  --o-filtered-table ./contigs/kraken2-autofmt-presence-absence-contigs.qza

Alpha Diversity#

Here we’ll examine and compare the community richness between our control and treatment groups.

Observed Features#

To start with, we’ll generate an observed features vector from our presence/absence feature table:

qiime diversity alpha \
    --i-table ./contigs/kraken2-autofmt-presence-absence-contigs.qza \
    --p-metric "observed_features" \
    --o-alpha-diversity ./contigs/obs-features-autofmt-contigs.qza

Linear Mixed Effects#

We will use a linear mixed-effects model in order to manage the repeated measures in our dataset.

qiime longitudinal linear-mixed-effects \
  --m-metadata-file sample-metadata.tsv ./contigs/obs-features-autofmt-contigs.qza \
  --p-state-column day-relative-to-fmt \
  --p-group-columns autoFmtGroup \
  --p-individual-id-column PatientID \
  --p-metric "observed_features" \
  --o-visualization ./contigs/lme-obs-features-treatmentVScontrol-contigs.qzv

Beta Diversity#

Now that we better understand community richness trends, let’s look at differences in microbial composition.

Jaccard Distance Matrix PCoA creation#

Let’s first create the Jaccard distance matrix!

qiime diversity beta \
  --i-table ./contigs/kraken2-autofmt-presence-absence-contigs.qza \
  --p-metric jaccard \
  --o-distance-matrix ./contigs/jaccard-autofmt-contigs.qza

Now, let’s generate a PCoA from our Jaccard matrix.

qiime diversity pcoa \
  --i-distance-matrix ./contigs/jaccard-autofmt-contigs.qza \
  --o-pcoa ./contigs/jaccard-autofmt-pcoa-contigs.qza

Emperor Plot Creation#

Now that we have our Jaccard diversity PCoA, let’s visualize it!

qiime emperor plot \
  --i-pcoa ./contigs/jaccard-autofmt-pcoa-contigs.qza \
  --m-metadata-file sample-metadata.tsv \
  --o-visualization ./contigs/jaccard-autofmt-emperor-contigs.qzv

We can make week-relative-to-fmt a custom axis in our PCoA. This allows us to look at changes in microbial composition over the course of the study.

qiime emperor plot \
  --i-pcoa ./contigs/jaccard-autofmt-pcoa-contigs.qza \
  --m-metadata-file sample-metadata.tsv \
  --p-custom-axes week-relative-to-fmt \
  --o-visualization ./contigs/jaccard-autofmt-emperor-custom-contigs.qzv

Taxa Barplot Creation#

We will now explore the microbial composition in our contig-based analysis by visualizing a taxonomic bar plot. Note that we are using a FeatureTable[PresenceAbsence], and thus are not talking about relative abundance.

qiime taxa barplot \
  --i-table ./contigs/kraken2-autofmt-presence-absence-contigs.qza \
  --i-taxonomy ./contigs/kraken2-taxonomy-contigs.qza \
  --m-metadata-file sample-metadata.tsv \
  --o-visualization ./contigs/taxa-bar-plot-autofmt-contigs.qzv

Functional Analysis#

qiime moshpit eggnog-diamond-search \
  --i-sequences "./moshpit_tutorial/cache:contigs" \
  --i-diamond-db "./moshpit_tutorial/cache:eggnog_diamond_full"\
  --p-num-cpus 14 \
  --p-db-in-memory \
  --o-eggnog-hits "./moshpit_tutorial/cache:diamond_hits_contigs" \
  --o-table "./moshpit_tutorial/cache:diamond_feature_table_contigs" \
  --verbose

Here we will perform functional annotation of contigs to capture gene diversity!

Obtaining your Feature Table#

You know the drill by now ;)

wget -O ./contigs/eggnog-presence-absence-contigs.qza https://polybox.ethz.ch/index.php/s/YSXu2AaOFeusgRY/download

Feature Table Filtering#

Same here ;)

qiime feature-table filter-samples \
  --i-table ./contigs/eggnog-presence-absence-contigs.qza \
  --m-metadata-file sample-metadata.tsv \
  --p-where 'autoFmtGroup IS NOT NULL' \
  --o-filtered-table ./contigs/filtered-eggnog-presence-absence-contigs.qza

Jaccard Distance Matrix PCoA creation for gene diversity#

We will again start by calculating our Jaccard beta diversity matrix.

qiime diversity beta \
  --i-table ./contigs/filtered-eggnog-presence-absence-contigs.qza \
  --p-metric jaccard \
  --o-distance-matrix ./contigs/jaccard-diamond-autofmt-contigs.qza

Then, we will generate our PCoA from Jaccard matrix.

qiime diversity pcoa \
  --i-distance-matrix ./contigs/jaccard-diamond-autofmt-contigs.qza  \
  --o-pcoa ./contigs/jaccard-diamond-autofmt-pcoa-contigs.qza

Emperor Plot Creation for gene diversity#

Visualization time!

qiime emperor plot \
  --i-pcoa ./contigs/jaccard-diamond-autofmt-pcoa-contigs.qza \
  --m-metadata-file sample-metadata.tsv \
  --o-visualization ./contigs/jaccard-diamond-autofmt-pcoa-contigs.qzv

Breakout Session#

We’re going to take the next ~30 minutes or so to breakout into two groups and work on an interactive exercise! Each group will be examining the correlation for different beta diversity metrics. Group A will be examining the correlation between our Jaccard distance matrices for reads and contigs. Group B will be examining the correlation between our read-based Jaccard distance matrix and a read-based Bray-Curtis distance matrix.

For each group, please look through the instructions below - you’ll notice that we’re not providing you with explicit commands. This is a good time to do some problem solving and sleuthing together as a group. Once your group has generated and examined your particular visualization, discuss amongst yourselves - do these results seem reasonable? Is this what you’d expect? Why or why not?

Group A#

In this group exercise we will be examining the correlation between Jaccard distance matrices for both reads and contigs. Note that while you’ll already have the Jaccard distance matrix for contigs, you’ll need to generate this for reads. We will not be providing commands for this, so do some poking around in the help text for commands within the diversity plugin to see what might make sense for comparing these two distance matrices. :)

Group B#

In this group exercise we will be examining the correlation between a read-based Jaccard distance matrix and a read-based Bray-Curtis distance matrix. Note that you’ll need to generate both of these distance matrices (as we’ve only generated Jaccard for contigs in the tutorial). We will not be providing commands for this, so do some poking around in the help text for commands within the diversity plugin to see what might make sense for comparing these two distance matrices. :)

A couple of helpful hints while working through these exercises:

Make sure you utilize the --help command (at both the plugin and command levels) to help you figure out what methods you’ll need to use to generate these distance matrices, as well as a visual correlation between a set of two distance matrices.
When running multi-line commands, you’ve seen the backslash format at the end of each line (visually splitting each parameter), which you can continue utilizing. You can also type everything out in a single line - the result will be the same, but you can pick which formatting feels preferable to you.

MAG-based analysis#

Let’s start binning our contigs into MAGs using various tools and methodologies!

Read mapping

We first need to index the contigs obtained in the assembly step and map the original reads to those contigs using that index. This read mapping can then be used by the contig binner to figure out which contigs originated from the same genome and put those together.

qiime assembly index-contigs \
    --i-contigs "./moshpit_tutorial/cache:contigs" \
    --p-seed 100 \
    --p-threads 64 \
    --p-verbose \
    --p-num-partitions 4 \
    --o-index "./moshpit_tutorial/cache:contigs_index" \
    --verbose

qiime assembly map-reads-to-contigs \
    --i-indexed-contigs "./moshpit_tutorial/cache:contigs_index" \
    --i-reads "./moshpit_tutorial/cache:reads_no_host" \
    --p-seed 100 \
    --p-threads 64 \
    --p-num-partitions 4 \
    --o-alignment-map "./moshpit_tutorial/cache:reads_to_contigs" \
    --verbose

Binning

Finally, we are ready to perform contig binning. This process involves categorizing contigs into distinct bins or groups based on their likely origin from different microbial species or strains within a mixed community. Here, we will use the MetaBAT 2 tool, which uses tetranucleotide frequency together with abundance (coverage) information to assign contigs to individual bins.

qiime moshpit bin-contigs-metabat \
    --i-contigs "./moshpit_tutorial/cache:contigs" \
    --i-alignment-maps "./moshpit_tutorial/cache:reads_to_contigs" \
    --p-seed 100 \
    --p-num-threads 128 \
    --p-verbose \
    --o-mags "./moshpit_tutorial/cache:mags" \
    --o-contig-map "./moshpit_tutorial/cache:contig_map" \
    --o-unbinned-contigs "./moshpit_tutorial/cache:unbinned_contigs" \
    --verbose

This step generated a couple artifacts:

mags.qza: these are our actual MAGS, per sample.
contig-map.qza: this is a mapping between MAG IDs and IDs of contigs which belong to a given MAG.
unbinned-contigs.qza: these are all the contigs that could not be assigned to any particular MAG. From here, we will focus on the mags.qza artifact.

MAGs QC with BUSCO

BUSCO is used here to assess the completeness and quality of MAGs by searching for single-copy orthologous genes within the genomes.

qiime moshpit evaluate-busco \
    --i-bins "./moshpit_tutorial/cache:mags" \
    --p-lineage-dataset bacteria_odb10 \
    --p-cpu 196 \
    --o-visualization "./moshpit_tutorial/results/mags.qzv" \
    --verbose

BUSCO QC#

Here we will have a look at our BUSCO results to assess the completeness and quality of MAGs!

wget -O busco-qc.qzv https://polybox.ethz.ch/index.php/s/fzAA003m6UVw5je/download

MAGs dereplication

Dereplication involves removing duplicate or nearly identical MAGs to reduce redundancy and improve downstream analyses. To dereplicate our MAGs, we will:

Compute hash sketches of every genome using sourmash - you can think of these sketches as tiny representations of our genomes (sourmash compresses a lot of information into much smaller space).
Compare all of those sketches (genomes) to one another to generate a matrix of pairwise distances between our MAGs.
Dereplicate the genomes using the distance matrix and a fixed similarity threshold. The last action will simply choose the longest genome from all of the genomes belonging to the same cluster, given a similarity threshold.

qiime sourmash compute \
    --i-sequence-file "./moshpit_tutorial/cache:mags" \
    --p-ksizes 35 \
    --p-scaled 10 \
    --o-min-hash-signature "./moshpit_tutorial/cache:mags_minhash" \
    --verbose

qiime sourmash compare \
    --i-min-hash-signature "./moshpit_tutorial/cache:mags_minhash" \
    --p-ksize 35 \
    --o-compare-output "./moshpit_tutorial/cache:mags_dist_matrix" \
    --verbose

qiime moshpit dereplicate-mags \
    --i-mags "./moshpit_tutorial/cache:mags" \
    --i-distance-matrix "./moshpit_tutorial/cache:mags_dist_matrix" \
    --p-threshold 0.99 \
    --o-dereplicated-mags "./moshpit_tutorial/cache:mags_derep" \
    --o-feature-table "./moshpit_tutorial/cache:mags_ft" \
    --verbose

MAGs taxonomic annotation workflow

This workflow focuses on annotating MAGs with taxonomic information using Kraken2, a tool for taxonomic classification. In this tutorial we will perfom taxonomic and functional annotation on dereplicated MAGs.

MAGs classify with Kraken2

MAGs are classified taxonomically using Kraken2, with parameters set for a confidence threshold and minimum base quality.

qiime moshpit classify-kraken2 \
    --i-seqs "./moshpit_tutorial/cache:derep_mags" \
    --i-kraken2-db "./moshpit_tutorial/cache:kraken_standard" \
    --p-threads 48 \
    --p-confidence 0.6 \
    --p-minimum-base-quality 20 \
    --p-num-partitions 4 \
    --o-reports "./moshpit_tutorial/cache:kraken_reports_derep_mags" \
    --o-hits "./moshpit_tutorial/cache:kraken_hit_derep_mags" \
    --verbose

Obtaining our Kraken2 reports#

QIIME 2 does not stop you from using your favorite tools with its output! First, let’s obtain an artifact containing Kraken 2 annotated MAGs from this dataset. We will visualize some of them with pavian.

wget -O kraken2-reports-mags-derep.qza https://polybox.ethz.ch/index.php/s/n0L2vm16C1J6MHe/download

Kraken2 annotation reports export#

Now, let’s export this QIIME artifact and explore!

qiime tools export \
  --input-path kraken2-reports-mags-derep.qza \
  --output-path kraken2-reports-mags-derep

Provenance Replay#

Without looking back through all of the commands we ran in this tutorial, how many of you feel confident that you could re-run our analyses from memory? If you don’t feel confident in that, you’re not alone! It is very common to have difficulty remembering the exact commands you ran for a past analysis (or trying to figure out the commands that someone else ran from an external analysis). Even if you write down all of the steps you’ve taken, humans are fallible and our memories aren’t perfect.

Each QIIME 2 Result (i.e. Artifact or Visualization) contains provenance that can be used as a reference to reconstruct the commands that were used to generate said result. Let’s take a look at the provenance associated with one of the visualizations from our read-based analysis as an example!

While using provenance to manually reconstruct the commands used to generate a result is a reasonable workflow for one or two results, we need a more automated solution to reconstruct the commands for a larger analysis - such as the analyses we ran in this workshop. Luckily, provenance replay can handle this for us!

We’ll start off by running provenance replay for all of the read-based results we’ve generated during this tutorial. We can run provenance replay on the entire reads directory. This will provide us with a replay supplement that contains all of the upstream commands used to generate each result, any relevant citations associated with each of the commands used (in BibTex format), and the recorded metadata used in each command.

qiime tools replay-supplement \
--in-fp ./reads \
--out-fp reads-replay-output

On your own, try generating the replay supplement for all of the contig-based results, and reconstruct a few of the commands used in that analysis!

Metagenomics with QIIME 2 - Spring 2024 Leiden & Zurich Tutorial

Contents

Metagenomics with QIIME 2 - Spring 2024 Leiden & Zurich Tutorial#

Setup#

Directory structure#

Metadata#

Read-based analysis#

Obtaining your Feature Table and Taxonomy Table#

Feature Table Filtering#

Taxa Barplot Creation#

Differential Abundance#

Feature Table Preparation#

Differential Abundance with ANCOM-BC#

Contig-based analysis#

QUAST QC#

Obtaining your Feature Table and Taxonomy Table#

Feature Table Filtering#

Alpha Diversity#

Observed Features#

Linear Mixed Effects#

Beta Diversity#

Jaccard Distance Matrix PCoA creation#

Emperor Plot Creation#

Taxa Barplot Creation#

Functional Analysis#

Obtaining your Feature Table#

Feature Table Filtering#

Jaccard Distance Matrix PCoA creation for gene diversity#

Emperor Plot Creation for gene diversity#

Breakout Session#

Group A#

Group B#

MAG-based analysis#

BUSCO QC#

Obtaining our Kraken2 reports#

Kraken2 annotation reports export#

Provenance Replay#