Skip to contents

Rarefying Phyloseq Data

Source: R/rarefying.R
rarefying.Rd

This function performs rarefaction on a normalized phyloseq object. Rarefaction is based on the biomass of each sample to identify the minimum sampling depth across the dataset. By rarefying all samples to this sampling depth, we ensure that the data are normalized for sequencing effort. This step prevents an overestimation of genus abundance in samples with higher sequencing depths, allowing for more accurate and comparable results.

Usage

rarefying(physeq = physeq, norm_method = NULL, iteration = 100)

Arguments

physeq

A phyloseq object containing the microbiome data. This is the input object that the function processes.

norm_method

A character string specifying the normalization method. Acceptable values are:

  • NULL: Use this option if no FCM or qPCR data is available, or if you wish to retain only relative abundances.

  • "fcm": Use this option if the data have been normalized using flow cytometry (FCM).

  • "qpcr": Use this option if the data have been normalized using quantitative PCR (qPCR).

Value

The rarefied phyloseq object is returned and also saved as an .rds file. The file name and location depend on the specified normalization method:

  • For "fcm": "project_name_phyloseq_asv_level_fcm_normalised_cell_concentration_rarefied.rds"

  • For "qpcr": "project_name_phyloseq_asv_level_qpcr_normalised_cell_concentration_rarefied.rds"

Details

  • For "fcm" normalization:

    • Rarefies only the fcm normalized data, while the copy number corrected data remains unchanged.

    • Rarefies based on the calculated sampling depth, derived from the ratio of total reads per sample to the estimated cell counts.

    • Uses a custom function (avgrarefy) to perform multiple iterations of rarefaction and averages the results.

  • For "qpcr" normalization:

    • Rarefies only the qpcr normalized data, while the copy number corrected data remains unchanged.

    • Rarefies based on the calculated sampling depth, derived from the ratio of total reads per sample to predicted 16S rRNA gene copy numbers.

    • Uses the same avgrarefy function for averaging rarefied counts.

  • The input values are subject to a scaling limit of 1e7. If input values exceed this limit due to the prior normalization, all values are scaled down by a calculated scaling factor. The rarefied counts are rescaled back to their original scale after rarefaction. However, due to this scaling and the rounding of scaled values, slight variations in the rarefied counts may occur.

The function uses parallel processing to improve the efficiency of rarefaction. It saves the rarefied phyloseq object as an .rds file in the appropriate output folder.

Note

Ensure that the phyloseq object is properly normalized before applying this function. Missing or invalid rarefy_to values will result in warnings and skipped samples.

Examples

# Rarefy using FCM normalization
rarefied_physeq <- rarefying(physeq = normalised_physeq, norm_method = "fcm")
#> Error in log_message(paste("Step 9: rarefied data", paste(projects, collapse = ", ")),     log_file): could not find function "log_message"

# Rarefy using qPCR normalization
rarefied_physeq <- rarefying(physeq = normalised_physeq, norm_method = "qpcr")
#> Error in log_message(paste("Step 9: rarefied data", paste(projects, collapse = ", ")),     log_file): could not find function "log_message"