Importing Data into MicrobiomeDB

library(MicrobiomeDB)
#> Warning: replacing previous import 'S4Arrays::makeNindexFromArrayViewport' by
#> 'DelayedArray::makeNindexFromArrayViewport' when loading 'SummarizedExperiment'

Importing Data

The MicrobiomeDB package supports importing data from the following formats:

1. TreeSummarizedExperiment
2. phyloseq
3. BIOM
4. QIIME2 output files
5. Mothur output files
6. dada2 output files
7. HUMAnN output files

For formats other than TreeSummarizedExperiment the mia package is also required. You can install it like:

if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")

BiocManager::install("mia")

Please note that Linux and MacOS with source-level installations require the ‘gsl’ system dependency for mia.

On Debian or Ubuntu

sudo apt-get install -y libgsl-dev

On Fedora, CentOS or RHEL

sudo yum install libgsl-devel

On macOS (source installations are not common on macOS, so this step is not usually necessary)

brew install gsl

TreeSummarizedExperiment

The MicrobiomeDB package natively supports importing TreeSummarizedExperiment objects. There is a function importTreeSummarizedExperiment (as well as the alias importTreeSE) which will produce an MbioDataset object from a TreeSummarizedExperiment. It is used in the following way:

data(GlobalPatterns, package="mia")
tse <- GlobalPatterns

## no normalization, with raw values
mbioDataset <- importTreeSummarizedExperiment(
      tse, 
      normalizationMethod = "none", 
      keepRawValues = TRUE, 
      verbose = TRUE
)

## TSS normalization, drop raw values
mbioDataset <- importTreeSummarizedExperiment(
      tse, 
      normalizationMethod = "TSS", 
      keepRawValues = FALSE, 
      verbose = TRUE
)

## TSS normalization, keep raw values
mbioDataset <- importTreeSummarizedExperiment(
      tse, 
      normalizationMethod = "TSS", 
      keepRawValues = TRUE, 
      verbose = TRUE
)

Phyloseq

First, make sure phyloseq is installed:

if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")

BiocManager::install("phyloseq")

Then, in order to import phyloseq objects, the mia package is also required. Once that is installed (see above), you can do the following:

data(GlobalPatterns, package="phyloseq")

mbioDataset <- importPhyloseq(
      normalizationMethod = "none", 
      keepRawValues = TRUE, 
      verbose = TRUE, 
      GlobalPatterns
)

BIOM

First, make sure biomformat is installed:

if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")

BiocManager::install("biomformat")

Then, in order to import biom files and objects, the mia package is also required. Once that is installed (see above), you can do the following:

## passing the file name directly
rich_dense_file = system.file("extdata", "rich_dense_otu_table.biom",
                              package = "biomformat")

mbioDataset <- importBIOM(
      normalizationMethod = "none", 
      keepRawValues = TRUE, 
      verbose = TRUE, 
      rich_dense_file
)

## passing a file already read into R as a biom object
rich_dense_biom = biomformat::read_biom(rich_dense_file)

mbioDataset <- importBIOM(
      normalizationMethod = "none", 
      keepRawValues = TRUE, 
      verbose = TRUE, 
      rich_dense_biom
)

QIIME2

In order to import QIIME2 output files, the mia package is also required. Once that is installed (see above), you can do the following:

featureTableFile <- system.file("extdata", "table.qza", package = "mia")
taxonomyTableFile <- system.file("extdata", "taxonomy.qza", package = "mia")

mbioDataset <- importQIIME2(
      normalizationMethod = "none", 
      keepRawValues = TRUE, 
      verbose = TRUE, 
      featureTableFile, 
      taxonomyTableFile
)

Mothur

In order to import Mother output files, the mia package is also required. Once that is installed (see above), you can do the following:

counts <- system.file("extdata", "mothur_example.shared", package = "mia")
taxa <- system.file("extdata", "mothur_example.cons.taxonomy", package = "mia")
meta <- system.file("extdata", "mothur_example.design", package = "mia")

mbioDataset <- importMothur(
      normalizationMethod = "none", 
      keepRawValues = TRUE, 
      verbose = TRUE, 
      counts, 
      taxa, 
      meta
)

dada2

First, make sure dada2 is installed:

if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")

BiocManager::install("dada2")

Then, in order to import dada2 output files, the mia package is also required. Once that is installed (see above), you can do the following:

fnF <- system.file("extdata", "sam1F.fastq.gz", package="dada2")
fnR = system.file("extdata", "sam1R.fastq.gz", package="dada2")
dadaF <- dada2::dada(fnF, selfConsist=TRUE)
#> Initializing error rates to maximum possible estimate.
#> selfConsist step 1 .
#>    selfConsist step 2
#>    selfConsist step 3
#>    selfConsist step 4
#> Convergence after  4  rounds.

dadaR <- dada2::dada(fnR, selfConsist=TRUE)
#> Initializing error rates to maximum possible estimate.
#> selfConsist step 1 .
#>    selfConsist step 2
#>    selfConsist step 3
#>    selfConsist step 4
#> Convergence after  4  rounds.

mbioDataset <- importDADA2(
      normalizationMethod = "none", 
      keepRawValues = TRUE, 
      verbose = TRUE, 
      dadaF, 
      fnF, 
      dadaR, 
      fnR
)

HUMAnN

In order to import HUMAnN output files, the mia package is also required. Once that is installed (see above), you can do the following:

## this is the merged output file
file_path <- system.file("extdata", "humann_output.tsv", package = "mia")

mbioDataset <- importHUMAnN(
      normalizationMethod = "none", 
      keepRawValues = TRUE, 
      verbose = TRUE, 
      file_path
)