WSSmod

WSSmod provides tools to construct Weighted Sum of Scores (WSS) module-level summaries from proteomic data. Proteins are grouped into STRING network modules, sign-flipped to align with the direction of core biomarker association, and combined into a single module score per sample using a weighted average based on network connectivity (alpha centrality) and biomarker association strength.

This package accompanies the Longo Lab plasma biomarker proteomics work.

Installation

You can install the development version of WSSmod from GitHub with:

# install.packages("pak")
pak::pak("Longo-Lab/WSSmod")

Example

Using a custom module/weighting table:

library(WSSmod)

expr_matrix <- matrix(
  c(1, 2, 3, 4, 5, 6, 7, 8),
  nrow = 2,
  dimnames = list(c("sample1", "sample2"), c("geneA", "geneB", "geneC", "geneD"))
)

result <- data.table::data.table(
  module = c("mod1", "mod1", "mod1", "mod2"),
  symbol = c("geneA", "geneB", "geneC", "geneD"),
  mean_beta = c(1, -1, 1, 1),
  mean_alpha_scaled = c(1, 1, 2, 1)
)

WSSmod::calculate_WSS(expr_matrix, result, min_module_size = 1)
#> $scores
#>         mod1 mod2
#> sample1  2.0    7
#> sample2  2.5    8
#> 
#> $summary_stats
#>    module   Min    Q1 Median  Mean    Q3   Max
#>    <char> <num> <num>  <num> <num> <num> <num>
#> 1:   mod1     2 2.125   2.25  2.25 2.375   2.5
#> 2:   mod2     7 7.250   7.50  7.50 7.750   8.0
#> 
#> $module_sd
#>      mod1      mod2 
#> 0.3535534 0.7071068 
#> 
#> $module_proteins
#> $module_proteins$mod1
#> [1] "geneC" "geneA" "geneB"
#> 
#> $module_proteins$mod2
#> [1] "geneD"

WSSmod also ships prebuilt module/weighting reference sets so you don’t have to build your own module table from scratch. calculate_WSS() uses prebuilt = "core_AD_plasma_biomarkers" by default when result = NULL:

list_prebuilt_wss()

# expr_matrix here has samples as rows and protein/gene symbols as columns,
# matching the symbols used by the prebuilt reference set
WSSmod::calculate_WSS(expr_matrix)

# Human-readable module term labels for the same reference set
wss_prebuilt_terms("core_AD_plasma_biomarkers")

Predicting module scores from core biomarkers

predict_WSS() predicts module scores for a prebuilt reference set from a small set of covariates, using a fitted joinet model, without needing the full proteomic panel. For "core_AD_plasma_biomarkers" this predicts all 75 module scores from Age, Gender, and the 8 core plasma biomarkers used in the original analysis. This requires the glmnet and joinet packages.

The 8 biomarker columns must already be rank-based inverse-normal transformed before calling predict_WSS(). normalize_wss_biomarkers() builds these columns (keeping the same column names) from raw biomarker values using one of two methods:

• method = "project" (default) — project each raw value onto a bundled reference distribution via project_rank_norm(). Works for a single new patient or a small/differently-distributed cohort, since each value is scored independently against the fixed reference.
• method = "self" — rank-normalize (RNOmni::RankNorm()) within your own batch, if you have a sizeable cohort of your own with a distribution you’re comfortable normalizing against directly instead of the bundled reference. Requires multiple samples, since a rank transform needs multiple values to rank against.

# raw_biomarkers: one row per sample, the 8 raw (un-normalized) biomarker columns.
# Works for a single patient (method = "project" is the default):
raw_biomarkers <- data.frame(
  PlasmaPTau181 = 1.5, PlasmaAB142P = 25, PlasmaAB140P = 300, PlasmaABRatio = 0.09,
  PlasmapTau217 = 0.3, PlasmapTau217_AB42Ratio = 0.012, PlasmaGFAP = 60, PlasmaNfL = 22
)
normalized <- normalize_wss_biomarkers(raw_biomarkers)

newdata <- cbind(Age = 72, Gender = 1, normalized) # Gender: 1 = male, 0 = not male
pred <- predict_WSS(newdata)
pred$meta