the whole multi-omics GSVA (moGSVA) pipeline: per-omics enrichment scoring,<br /> cross-omics weighted fusion with a consistency reward, and a gene-to-pathway<br /> permutation test for significance assessment.
mogsva_opts(
score.method = gsva,
kcdf = Gaussian,
fusion.strategy = Equal,
consistency.lambda = 0.25,
balance.feature.size = FALSE,
min.size = 5,
max.size = 500,
ssgsea.alpha = 0.25,
ssgsea.normalize = TRUE,
permutations = 0,
permutation.method = ssgsea,
permutation.pathways = NULL,
permutation.gene.limit = 0,
max.permutation.pathways = 40,
seed = 20261026,
compute.pooled = TRUE,
verbose = TRUE);
score.method
The enrichment scoring method applied independently inside each omics layer (recommended Methods.gsva, the default). Other choices are ssGSEA, z-score and PLAGE. [as Methods]
kcdf
The kernel density estimation type used by the GSVA scoring step: KCDFs.Gaussian for continuous data (default), KCDFs.Poisson for count data, or KCDFs.none for the empirical cumulative distribution function. [as KCDFs]
fusion.strategy
The cross-omics weighted fusion strategy. Defaults to FusionMethods.Equal (a plain average, i.e. baseline 1). Alternatives: variance weighting, information-entropy weighting, mutual-information consistency weighting and a balanced scheme that also applies the per-omics feature-count balancing term. [as FusionMethods]
consistency.lambda
The hyper-parameter λ (lambda) of the cross-omics consistency reward term. A value of 0 disables the reward and reduces the fusion to a pure weighted average. Default 0.25. [as double]
balance.feature.size
Whether to superimpose the per-omics feature-count balancing term 1/log2(2+|S_k|) so that omics layers with many mapped members cannot dominate the fused score. Off by default so that the equal-weight fusion stays strictly identical to the simple per-omics average. [as boolean]
min.size
Minimum gene-set size (number of mapped features) after the feature mapping step. Gene sets smaller than this are dropped. Default 5. [as integer]
max.size
Maximum gene-set size (number of mapped features) after the feature mapping step. Gene sets larger than this are dropped. Default 500. [as integer]
ssgsea.alpha
The α (alpha) parameter of the ssGSEA scoring path. Default 0.25. [as double]
ssgsea.normalize
Whether the ssGSEA scoring path normalizes the enrichment scores. Default TRUE. [as boolean]
permutations
Number of permutations for the gene-to-pathway permutation test. 0 (the default) skips the significance assessment entirely. [as integer]
permutation.method
The scoring method used to build the permutation null distribution. Defaults to ssGSEA (rank based) to avoid repeating the expensive kernel density estimation. [as Methods]
permutation.pathways
A subset of pathway identifiers to evaluate in the permutation test. When nothing is supplied the first max_permutation_pathways pathways from the fused result are used. [as string]
permutation.gene.limit
Upper bound on the number of features kept in the permutation view (0 means no trimming). When trimming is applied, the target pathway members are kept first and the remaining budget is filled with background features using a fixed seed. [as integer]
max.permutation.pathways
Maximum number of pathways evaluated in the permutation test when no explicit pathway subset is supplied. Default 40. [as integer]
seed
Random seed used for deterministic data derivation and permutation shuffling so that runs are fully reproducible. Default 20261026. [as integer]
compute.pooled
Whether to also compute the ablation control "concatenate then single-omics GSVA" baseline. Default TRUE. [as boolean]
verbose
Whether to emit progress/diagnostic logging. Default TRUE. [as boolean]
env
The R# runtime environment (used for error reporting). [as Environment]