noriakis/q2-pathway

QIIME 2 plugin for analyzing biological pathway information based on gene family abundances

q2-pathway

QIIME 2 plugin for analyzing gene family abundances and biological pathway information obtained from 16S rRNA gene sequencing data.

Installation

After activating the QIIME 2 environment:

mamba install -c noriakisato -c bioconda q2-pathway

ALDEx2 and DESeq2 shuold be installed in the QIIME 2 environment if you are to use the statistics from the packages:

install.packages("BiocManager")
BiocManager::install("ALDEx2")
BiocManager::install("DESeq2")

If kegg module is to be used, pykegg must be installed:

pip install pykegg

q2-pathway

This plugin is used to analyze the functional prediction results from 16S rRNA gene sequencing dataset and optionally the profile from shotgun metagenomes.

infer module can perform an inferrence based on Piphillin or Tax4Fun2 algorithm. The Tax4Fun2 software and algorithm will be downloaded and installed at the first command launch of infer_t4f2 (the original files can be found here under GNU General Public License v3.0 or later).

For Piphillin, the databases can be found here (RefSeq, NCBI usage policies) and here (GTDB, CC BY-SA 4.0).

Suppose that rep-seqs.qza is an artifact containing the representative sequence (of ASV or OTU) and table-dada2.qza contains the count table of the sequences:

## This runs the Piphillin algorithm
qiime pathway infer \
    --i-sequences rep-seqs.qza \
    --i-seq-table table-dada2.qza \
    --i-database piphillin_refseq.qza \
    --o-table infer_piphillin.qza \
    --p-threads 0

## This runs the Tax4Fun2 algorithm
qiime pathway infer-t4f2 \
    --i-sequences rep-seqs.qza \
    --i-seq-table table-dada2.qza \
    --o-table infer_t4f2.qza

Also, gsea module is implemented for performing GSEA using fgsea, based on the KEGG PATHWAY mapping. The users should perform with --verbose for inspecting the GSEA output (like the existing of the ties). Althoug there is already a plugin (q2-aldex2), the function can rank the genes based on the statistics from ALDEx2. One should install ALDEx2 (BiocManager::install("ALDEx2")) beforehand.

## This will run GSEA based on the log2 fold changes calculated from DESeq2
qiime pathway gsea \
    --i-ko-table ko_metagenome.qza \
    --m-metadata-file metadata.tsv \
    --o-visualization gsea_output \
    --p-method deseq2

The summarize module reports and compares the gene family abundance table between the tables produced by multiple inference methods including q2-picrust2. Using q2-sapienns, the results from the shotgun metagenomics data can also be compared. In that case, the converting table for shotgun accession to 16S accession (Metadata containing converted column) must be provided to --p-convert-table. The correlation metrics can be chosen from spearman, pearson, kendall by --p-method. Also, the correlation based on the p-values proposed in Sun et al. 2020. can be calculated by specifying --p-use-p. Metadata can be omitted.

qiime pathway summarize \
    --i-tables ko_metagenome.qza \
    --i-tables ko_metagenome.qza \
    --m-metadata-file metadata.tsv \
    --o-visualization vis_output

Finally, the kegg module is implemented for visualization of KEGG PATHWAY images colored by the statistics calculated from comparing the categorical variables in the metadata. The plugin uses pykegg to produce the images.

qiime pathway kegg \
    --i-tables ko_metagenome.qza \
    --m-metadata-file metadata.tsv \
    --p-pathway-id ko00240 \
    --o-visualization pathway_output

Bugs and errors

If you find bugs, suggestions, or errors, please kindly report them to Issues, or make a pull request, or report it directly to e-mail.