FROGSSTAT DESeq2 Preprocess

Context

Are there ASV with differential abundance between 2 conditions ? And which are they ?
To answer these questions, we perform a differential abundance analysis using DESeq2 on the phyloseq object.

The package DESeq2 provides methods to test for differential expression by use of negative binomial generalized linear models

• 1^st step: FROGSSTAT DESeq2 Preprocess
• 2^nd step : FROGSSTAT DESeq2 Visualisation

Available to analyse ASV or Function (from FROGSFUNC tools)

1^st step: create dds object with FROGSSTAT DESeq2 Preprocess

FROGSSTAT DESeq2 Preprocess launches Rscript to compute differential abundance analysis using DESeq2 on a Phyloseq object.

use cases:

• for ASVs
  • FROGSSTAT DESeq2 Preprocess needs the phyloseq object (.Rdata) created by FROGSSTAT Phyloseq Import Data
  • and
  • creates Deseq2 object (dds)
• for functions
  • FROGSSTAT DESeq2 Preprocess creates the phyloseq object (.Rdata)
  • and
  • creates Deseq2 object (dds)

Command line

usage: deseq2_preprocess.py [-h] [--debug] [--version] -v VAR -a
                            {ASV,FUNCTION} [-d DATA] [-f INPUT_FUNCTIONS]
                            [-s SAMPLEFILE] [--out-Phyloseq OUT_PHYLOSEQ]
                            [-o OUT_RDATA] [-l LOG_FILE]

Launch Rscript to generate dataframe of DESEq2 from a phyloseq object in RData
file

optional arguments:
  -h, --help            show this help message and exit
  --debug               Keep temporary files to debug program.
  --version             show programs version number and exit
  -v VAR, --var VAR     Experimental variable suspected to have an impact on
                        abundances. You may precise complexe string such as
                        variables with confounding effect (ex:
                        Treatment+Gender or Treatmet*Gender)

Inputs:
  -a {ASV,FUNCTION}, --analysis {ASV,FUNCTION}
                        Type of data to perform the differential analysis.
                        ASV: DESeq2 is run on the ASVs abundances table.
                        FUNCTION: DESeq2 is run on FROGSFUNC function
                        abundances table (frogsfunc_functions_unstrat.tsv from
                        FROGSFUNC function step).

 ASV :
  -d DATA, --data DATA  The path of RData file containing a phyloseq object,
                        result of FROGS Phyloseq Import Data. Required.

 FUNCTION :
  -f INPUT_FUNCTIONS, --input-functions INPUT_FUNCTIONS
                        Input file of metagenome function prediction
                        abundances (frogsfunc_functions_unstrat.tsv from
                        FROGSFUNC function step). Required. (default: None).
  -s SAMPLEFILE, --samplefile SAMPLEFILE
                        path to sample file (format: TSV). Required.
  --out-Phyloseq OUT_PHYLOSEQ
                        path to store phyloseq-class object in Rdata file.
                        [Default: function_data.Rdata]

Outputs:
  -o OUT_RDATA, --out-Rdata OUT_RDATA
                        The path to store resulting dataframe of DESeq2.
                        [Default: None]
  -l LOG_FILE, --log-file LOG_FILE
                        This output file will contain several information on
                        executed commands.

Example of command line:

• for ASVs

./deseq2_preprocess.py \
 --data data.Rdata \
 --analysis ASV \
 --log-file deseq2_preprocess_ASV.log \
 --out-Rdata deseq2_preprocess_ASV.Rdata \
 --var EnvType

• for functions

./deseq2_preprocess.py \
 --samplefile chaillou.sample \
 --input-functions frogsfunc_functions_unstrat_EC.tsv \
 --analysis FUNCTION \
 --log-file deseq2_preprocess_func.log \
 --out-Rdata deseq2_preprocess_func.Rdata \
 --out-Phyloseq phyloseq_functions.Rdata \
 --var EnvType

Galaxy for ASV analyses

Outputs for ASV analyses

a asv_dds.Rdata file that is a DESeq2 dataset stored in Rdata file.

Galaxy for function analyses

For FUNCTION: two files are requiered

• the files that contains predicted function abundances, it is named frogsfunc_functions_unstrat.tsv and produced by FROGSFUNC_2_function tool.
• the file that must contain the metadata that characterise each sample i.e. the conditions of experiment with sample ID in the first column as:
  

Outputs for function analyses

1. a function_data.Rdata that contains information of data in one phyloseq object.
2. a function_dds.Rdata that is a DESeq2 dataset stored in Rdata file. This result will be one of the input of the FROGSSTAT DESeq Visualisation tool.