Diversity Analyses

We will go through some of the statistical analyses that Qiime2 can perform on the data.

Generate tree for phylogenetic diversity analysis

qiime phylogeny align-to-tree-mafft-fasttree \
  --i-sequences {REP-SEQS}.qza \
  --o-alignment {ALIGNED-REP-SEQS}.qza \
  --o-masked-alignment {MASKED-ALIGNED-REP-SEQS}.qza \
  --o-tree {UNROOTED-TREE}.qza \
  --o-rooted-tree {ROOTED-TREE}.qza

Alpha and beta diversity analyses

A more detailed description of the following statistical tests can be found on the Moving Pictures tutorial page

Determine appropriate sampling depth

It is necessary to set the rarefaction or sampling depth for the analyses (see here for an explanation). This is set with the --p-sampling-depth parameter in the core-metrics plugin. You can look at the feature table summary to examine the number of reads for each sample to help set the sampling depth:

VISUALISATION: feature table summary

Run core metrics

Now we will run the core-metrics-phylogenetic pipeline, which runs a range of alpha and beta diversity tests on the data. This utilises the tree produced in the previous step in order to measure diversity based on phylogenetic distance.

qiime diversity core-metrics-phylogenetic \
  --i-phylogeny {ROOTED-TREE}.qza \
  --i-table {COMBINED-TABLE}.qza \
  --p-sampling-depth {NUMBER} \
  --m-metadata-file sample-metadata.tsv \
  --output-dir {CORE-METRICS-RESULTS}

Several statistical and visual outputs are produced, including PCOA plots. Here is one:

VISUALISATION: bray-curtis emperor plot

VISUALISATION: nad bray-curtis emperor plot

Alpha diversity (within samples)

We will now test for associations between categories in the sample metadata file and alpha diversity data.

qiime diversity alpha-group-significance \
  --i-alpha-diversity {CORE-METRICS-RESULTS}/faith_pd_vector.qza \
  --m-metadata-file sample-metadata.tsv \
  --o-visualization {CORE-METRICS-RESULTS}/faith-pd-group-significance.qzv

qiime diversity alpha-group-significance \
  --i-alpha-diversity {CORE-METRICS-RESULTS}/evenness_vector.qza \
  --m-metadata-file sample-metadata.tsv \
  --o-visualization {CORE-METRICS-RESULTS}/evenness-group-significance.qzv

Here is the output from evenness group significance

VISUALISATION: Evenness group significance

Beta diversity (between samples)

Now we will test whether the distances between samples within a group are more similar to each other than they are to samples from other groups. The ‘group’ you will get from the sample metadata categories. You can examine the PCoA produced in the core-metrics analyses above to determine what will likely give a significant result.

qiime diversity beta-group-significance \
  --i-distance-matrix {CORE-METRICS-RESULTS}/unweighted_unifrac_distance_matrix.qza \
  --m-metadata-file sample-metadata.tsv \
  --m-metadata-column {METADATA-CATEGORY} \
  --o-visualization {CORE-METRICS-RESULTS}/unweighted-unifrac-{METADATA-CATEGORY}-significance.qzv \
  --p-pairwise

Here is a permanova result using ‘sample-type’:

VISUALISATION: permanova result

Further analyses

There are many kinds of analyses you can run on the data, such as time series, and Mantel distance. See the Diversity plugin page for more information.

Importing from other programs

You can see that the two main components that you need to run the diversity analyses is a frequency table and representative sequences. This means we should be able to import results from other other programs, such as OBITools, into Qiime2 to run the statistics. We will try this using the table and ZOTUs from the previous workshop. The only other file you will need is a sample metadata file, which can be as simple or complicated as you need.

The program biom, which comes bundled with the Qiime2 installation, is used to convert to its format:

Import frequency table

biom convert -i zotutab_10_changed.txt -o belarus.biom --to-hdf5

The table in biom format can then be imported into Qiime2:

qiime tools import \
  --input-path belarus.biom \
  --type 'FeatureTable[Frequency]' \
  --input-format BIOMV210Format \
  --output-path belarus_feature-table.qza

Import Rep seqs (ZOTUs)

qiime tools import \
  --type 'FeatureData[Sequence]' \
  --input-path zotus_10.fasta  \
  --output-path belarus_zotu_rep_seqs.qza

Generate table stats so you can determine –p-sampling-depth

qiime feature-table summarize \
  --i-table belarus_feature-table.qza \
  --o-visualization belarus_feat-table.qzv \
  --m-sample-metadata-file fish_samples.tsv

qiime feature-table tabulate-seqs \
  --i-data belarus_zotu_rep_seqs.qza \
  --o-visualization belarus_zotu_rep-seqs.qzv

from examining feature table vis, interactive sample detail, can set minimum to 20000

Run core-metrics on samples

You could run the core-metrics-phylogenetics in order to use diversity (depends on the gene used), but for the sake of this example, the regular core-metrics will suffice.

qiime diversity core-metrics \
  --i-table belarus_feature-table.qza \
  --p-sampling-depth 20000 \
  --m-metadata-file fish_samples.tsv \
  --output-dir belarus-core-metrics-results

Going further with analyses

For more advanced statistical analyses, and to produce publication-quality figures (and have more control over visual details), there is a plug-in called Qiime2R:

Qiime2R Github page

There is also a tutorial on the Qiime2 Forum:

Qiime2R Tutorial

Getting Started with Qiime2

introductory tutorial for qiime2