OlinkAnalyze (version 4.2.0)

olink_normalization_qs: Quantile smoothing normalization of all proteins between two NPX projects.

Description

This function uses bridge samples to map quantiles of the non-reference dataset to the ones of the reference dataset. Mapped quantiles are used to transform the quantifications of the the non-reference dataset to the reference.

Usage

olink_normalization_qs(
  lst_df,
  ref_cols,
  not_ref_cols,
  bridge_samples,
  ref_product
)

Value

A "tibble" of Olink data in long format containing both input datasets with the quantile normalized quantifications.

Arguments

lst_df

A named list of the 2 input datasets. First element should be the reference dataset from Olink Explore HT and the second element should originate from Olink Explore 3072. (required)

ref_cols

A named list with the column names to use. Exported from olink_norm_input_check. (required)

not_ref_cols

A named list with the column names from the non-reference dataset. Exported from olink_norm_input_check. (required)

bridge_samples

Character vector of samples to be used for the quantile mapping. (required)

ref_product

Name of reference product. Must be one of "HT" or "Reveal".

Author

Amrita Kar Marianne Sandin Masoumeh Sheikhi Klev Diamanti

Details

In the case when a study is separated into multiple projects, an additional normalization step is needed to allow the data to be comparable across projects. Across different Olink products, some of the assays exist in corresponding but distinct NPX spaces. For those assays, the median of paired differences is insufficient for bridging as it only considers one anchor point (the median/50% quantile). Instead, quantile smoothing (QS) using multiple anchor points (5%, 10%, 25%, 50%, 75%, 90% and 95% quantiles) is favored to map the Explore 3072 data to the Explore HT distribution. The olink_normalization_qs() performs quantile smoothing bridging normalization between datasets from two Olink products (for example Olink Explore 3072 and Olink Explore HT) by performing the following steps:

• An empirical cumulative distribution function is used to map datapoints for the bridging samples from one product to the equivalent space in the other product.

• A spline regression model is constructed using unmapped and mapped data from one product, using anchor points from the quantiles defined above.

• The spline regression model is used to predict the normalized NPX values for all datapoints

More information on quantile smoothing and between product normalization can be found in the Bridging Olink Explore 3072 to Olink Explore HT tutorial.

Examples

# \donttest{
# Bridge samples
bridge_samples <- intersect(
  x = unique(OlinkAnalyze:::data_ht_small$SampleID),
  y = unique(OlinkAnalyze:::data_3k_small$SampleID)
) |>
  (\(x) x[!grepl("CONTROL", x)])()

# Run the internal function olink_norm_input_check
check_norm <- OlinkAnalyze:::olink_norm_input_check(
  df1 = OlinkAnalyze:::data_ht_small,
  df2 = OlinkAnalyze:::data_3k_small,
  overlapping_samples_df1 = bridge_samples,
  overlapping_samples_df2 = NULL,
  df1_project_nr = "P1",
  df2_project_nr = "P2",
  reference_project = "P1",
  reference_medians = NULL
)

# Named list of input datasets
lst_df <- list(
  check_norm$ref_df,
  check_norm$not_ref_df
)
names(lst_df) <- c(check_norm$ref_name, check_norm$not_ref_name)

ref_cols <- check_norm$ref_cols
not_ref_cols <- check_norm$not_ref_cols

qs_result <- OlinkAnalyze:::olink_normalization_qs(
 lst_df = lst_df,
 ref_cols = ref_cols,
 not_ref_cols = not_ref_cols,
 bridge_samples = bridge_samples,
 ref_product = "HT"
)
# }