NEWS

---

enrichit 0.2.6 (2026-10-06)

• fix permutation-based GSEA p-values in the fixed and adaptive engines (method = "sample" / "permute") for the default two-sided (scoreType = "std") calculation: the numerator counted only same-sign permutations while the denominator divided by all permutations, so p-values were too small under a symmetric null; the denominator now counts only permutations on the same side of zero as the observed ES, following the main fgsea convention (2026-10-06, Tue, closes #3)
• update the RcppEigen sparse-matrix interface in rwr_eigen_cpp() from MappedSparseMatrix to Map<SparseMatrix> for Eigen 5 compatibility (2026-10-06, Tue, PR #2)
• add a reportNA argument to gsea_gson() so pathways with NA multilevel p-values can be retained when needed (2026-10-04, Sun, closes DOSE#88)
• fix GSEA leading-edge edge cases by standardizing the rank = 0L sentinel for zero-overlap, N_R == 0, and non-finite running-score paths; align as_gseaResult() with the same contract; document the gsea_leading_edge_details() return contract; and tighten converter tests for the rank contract (2026-10-04–2026-10-05, Sun–Mon, closes DOSE#46)
• keep qvalue as NA when estimation fails instead of falling back to p.adjust, document the pvalue / p.adjust / qvalue semantics in as_enrichResult() and as_gseaResult(), and update converter tests accordingly (2026-10-05, Mon, addresses DOSE#24)

enrichit 0.2.5 (2026-09-24)

• ORA results now carry an oddsRatio column (Fisher's exact 2x2 odds ratio, placed next to FoldEnrichment). FoldEnrichment is a ratio of proportions, whereas the odds ratio is the effect size that the hypergeometric/Fisher test is built on; note it is the plain cross-product odds ratio and differs slightly from fisher.test()$estimate, which reports the conditional MLE (2026-09-22, Tue)
• report non-finite gene statistics instead of crashing with missing value where TRUE/FALSE needed: gsea_gson() checked is.unsorted() before validating the input, and a bare if (NA) aborted there, so an NA/NaN/Inf in the gene statistics never reached the finiteness check that explains the problem; gseaScores() now validates too and returns NA rather than aborting when the running score is not finite (2026-09-22, Tue)
• add generic constructors as_enrichResult() / as_gseaResult(): build 'enrichit' result objects from result tables of external enrichment tools (enrichr, g:Profiler, WebGestalt, fgsea, ...), with canonical column aliases, ID de-duplication, p-value validation, derived statistics (RichFactor / FoldEnrichment / zScore) when the query and background are known, and missing GSEA detail columns recomputed via gsea_leading_edge_details() (2026-09-21, Sun)

enrichit 0.2.4 (2026-09-17)

• stop emitting spurious no package '<...>' was found warnings when no input gene can be mapped (2026-09-17, Thu)
  • check_gene_id() routed its informational notices through yulab.utils::yulab_msg(), which builds the package citation banner and expects a package name; it therefore called packageDescription() on the notice text itself, so every run with unmappable input emitted three warnings such as no package '--> No gene can be mapped....' was found
  • the notices now go through message(), and yulab.utils::yulab_msg() is no longer imported
  • the sample of expected gene IDs shown in the notice no longer contains NA when a gene set has fewer than 100 genes

enrichit 0.2.3

• calibrate NSEA significance testing by switching the default to a whole-pipeline permutation null instead of GSEA's label-permutation test (2026-08-23, Sun)
  • the legacy nsea()/nsea_gson() pipeline ran GSEA's label-permutation test on the network-diffused scores; because network diffusion induces strong autocorrelation between neighbouring genes' scores, that test violates the exchangeability assumption of GSEA's permutation null and produces anti-conservative p-values (empirically ~2.5x the nominal false-positive rate)
  • significance = "whole_pipeline" (new default) builds the null distribution of the enrichment score by re-running the entire pipeline under the null: permute gene labels -> re-diffuse over the network with RWR -> recompute the enrichment score; p-values and NES are then derived from this null, automatically accounting for the smoothing induced by diffusion
  • significance = "internal" keeps the earlier internal label-permutation formulation on the diffused scores
• add significance, nPerm, seed and exponent arguments to nsea() and nsea_gson(); expose pvalueCutoff and pAdjustMethod in nsea_gson()
  • nPerm (default 1000) sets the number of whole-pipeline permutations; the smallest estimable p-value is 1/(nPerm + 1)
  • seed (default NULL) makes the whole-pipeline null reproducible; set a numeric seed for deterministic results
  • exponent (default 1) controls the weight of each step of the running enrichment score
  • nsea_gson() now applies pvalueCutoff to both the raw and the adjusted p-value (matching gsea_gson()), and accepts pAdjustMethod
• extend prepare_network() to accept a mechgraph object as well as an edge-list data.frame/matrix or a sparse matrix
  • the mechgraph edges table must have from/to columns; a numeric score or weight column, when present, supplies the edge weight, otherwise unit weights are used
  • the class is duck-typed by name, so enrichit does not hard-depend on the mechgraph package
• add a fast C++ kernel gsea_es_all_cpp() that computes the classic weighted enrichment score for many gene sets at once from the hit positions only, making it cheap enough to evaluate on every whole-pipeline permutation
• harden the NSEA internals
  • robust node-name extraction guards against Matrix accessor quirks on serialized sparse matrices
  • share single implementations of the RWR diffusion (.nsea_diffuse()) and TF-IDF specificity weighting (.nsea_specific_weights()); in "signed" mode positive and negative seeds are propagated in a single linear solve
• add regression tests for whole-pipeline reproducibility and for p-value calibration on null data

enrichit 0.2.2

• align gsea_gson() p-value filtering with the historical clusterProfiler/DOSE behavior: pvalueCutoff now requires both the raw p-value and the adjusted p-value (p.adjust) to pass the cutoff (previously only the raw p-value was filtered), restoring significant-pathway counts comparable to clusterProfiler <= 4.18.x (2026-08-14, Fri)
• clarify and harden the seed interface of GSEA for reproducibility (2026-08-14, Thu)
  • gsea() now treats seed = TRUE as a fixed default seed (consistent with the C++ default) instead of silently coercing it to the integer 1
  • gsea_gson() now exposes an explicit seed argument (previously only reachable through ...) and forwards it to gsea()
  • document that seed = FALSE (default) draws a fresh seed from R's RNG on each run, so results may vary between runs, while a numeric seed (or set.seed() before the call) makes the result reproducible; the C++ engine seeds its own RNG with this value
  • add regression tests asserting identical results across runs with a fixed seed

enrichit 0.2.1 (2026-08-04)

• fix gsea() to intersect gene sets with names(geneList) before applying minGSSize/maxGSSize, so the size filter constrains the actual overlap rather than the raw gene set size; also guard gsea_gson() against NA pvalue rows leaking into the result table (2026-08-04, Tue, clusterProfiler#824)

enrichit 0.2.0 (2026-07-01)

• extend multi-omics integration with pathway-level and topology-level workflows (2026-06-24, Wed)
  • add aggregate_enrichment() for multi-omics Late Fusion at the pathway level by aggregating multiple enrichResult/gseaResult objects
  • support correlation-aware p-value aggregation with method = "brown" in aggregate_omics()
  • support method = "weighted_mean" in aggregate_omics() for signed statistics with layer-specific weights
  • add mnsea() and mnsea_gson() for multi-layer network-based enrichment
  • add prepare_multilayer_network(), propagate_multilayer(), and collapse_multilayer_scores() for the multi-layer propagation pipeline
  • add mnseaResult to store multi-layer diffusion results, collapsed scores, layer weights, and cached explanation tables
  • precompute pathway-level and feature-level explanation caches inside mnseaResult
  • add get_mnsea_contribution() and extract_mnsea_subnetwork() for explanation-ready data extraction
  • provide the generic engine layer for downstream high-level wrappers such as clusterProfiler::mnseGO(), mnseKEGG(), mnseMKEGG(), and mnseWP()
• add aggregate_omics(), harmonize_ids() and select_features_for_ora() to support Multi-omics Early Integration
  • add conflict_policy parameter ("keep_all", "strict", "penalty") to handle directional conflicts in signed statistics (2026-06-23, Tue)
  • support early fusion before ora(), gsea(), and nsea() through a decoupled aggregation layer
  • these workflow helpers, together with aggregate_enrichment(), are designed to be reused by downstream packages for high-level multi-omics analysis
• add get_omics_contribution() and classify_omics_pattern() for Multi-omics contribution tracing
• implement Weighted Enrichment Analysis (2026-06-23, Tue)
  • add weight parameter to ora(), ora_gson(), gsea(), and gsea_gson()
  • support Weighted ORA using Wallenius' noncentral hypergeometric distribution via the BiasedUrn package
  • support Weighted GSEA by fusing external weights with ranked statistics
• implement Network-based Set Enrichment Analysis (NSEA) (2026-06-23, Tue)
  • add nsea() and nsea_gson() for network-ranked GSEA based on Random Walk with Restart (RWR)
  • add mode = "signed" support in nsea() and nsea_gson() for bidirectional network propagation using signed statistics
  • add prepare_network() for parsing and normalizing edge lists or sparse matrices
  • implement extremely fast RWR using RcppEigen sparse matrix multiplication
  • introduce zero-dependency integration strategy for network propagation followed by multilevel GSEA
  • provide the generic engine layer for downstream high-level wrappers such as clusterProfiler::nseGO(), nseKEGG(), nseMKEGG(), and nseWP()
• align multilevel GSEA rank scaling with fgsea::prepareStats() to reduce result drift relative to the long-used fgsea backend (2026-06-22, Mon)
  • replace the fixed * 1e6 scaling in prepare_gsea_inputs() with fgsea-style total-weight normalization and integer rounding
  • add a regression test that compares gsea(method = "multilevel") against fgsea::fgseaMultilevel() on the same ranked input
• exclude zero-overlap gene sets from ora_gson() before multiple-testing correction (2026-06-22, Mon)
  • keep Count = 0 rows out of p.adjust/qvalue so ORA results match historical DOSE/clusterProfiler behavior
  • resolves inflated adjustment in downstream clusterProfiler::enricher(), enrichKEGG(), and compareCluster() workflows (e.g. compound KEGG analyses, #821 & #819 of 'clusterProfiler')

enrichit 0.1.5 (2026-06-16)

• add Bayesian term selection for ORA results (2026-06-16, Tue)
  • add bayes_enrich() for estimating posterior probabilities of active explanatory terms from enrichResult objects
  • add bayes_summary() to return Bayesian enrichment results ordered by posterior probability
  • support candidate term spaces from top ranked result rows, significant as.data.frame() rows, all x@result rows, or explicit term IDs
  • keep the output as an enrichResult object with posterior, posterior odds, Bayesian rank, active flag, and covered-gene columns

enrichit 0.1.4 (2026-04-08)

• fix GSON helper functions (2026-04-06, Mon)
  • correct TERM2NAME() for GSON objects to map gsid to term names properly
  • correct TERMID2EXTID() for GSON objects to return gene vectors by requested term order
  • return character(0) for missing terms to keep downstream behavior stable
  • resolves malformed Description.* columns in downstream clusterProfiler::groupGO()

enrichit 0.1.3 (2026-03-10)

• fix gsea_gson() (2026-03-10, Tue)
  • force sorting of geneList to ensure result object is consistent with input
  • resolves issue with enrichplot::gseaplot showing incorrect metric/color alignment

enrichit 0.1.2 (2026-02-26)

• improve robustness of calculate_qvalue() (2026-02-02, Mon)
  • handle missing qvalue package gracefully
  • retry with different parameters if qvalue() fails
  • handle invalid p-values (NA, infinite, out of range)
• validate p-values in ora_gson() (2026-02-02, Mon)
  • ensure p-values are within [0, 1] range
  • warn and report invalid p-values
• optimize ORA p-value calculation in C++ (2026-02-02, Mon)
  • use phyper instead of summing dhyper for better performance and precision

enrichit 0.1.1 (2026-02-01)

• update setReadable() to support converting gene ID to other types (not limited to SYMBOL) (2026-01-21, Wed)
• add organism slot in compareClusterResult (2026-01-20, Tue)

enrichit 0.1.0 (2026-01-18)

• fix bugs in gsea_gson() and ora_gson() (2026-01-11, Sun)
  • handle missing columns (e.g., qvalues) gracefully by filling with NA
  • handle NA or duplicate gene set IDs in result rownames to prevent errors
• improve robustness of calculate_qvalue() (2026-01-11, Sun)
  • return NA instead of NULL when qvalue calculation fails
• update ora_gson() output columns (2026-01-11, Sun)

enrichit 0.0.9 (2026-01-10)

• add leading edge analysis for GSEA (2026-01-10, Sat)

enrichit 0.0.8 (2025-12-22)

• fixed bugs of multilevel GSEA in p value calculation (2025-12-10, Wed)
  • by learning the source code of 'fgsea'

enrichit 0.0.7

• add gseaScores function (2025-12-07, Sun)
  • to calculate GSEA scores for a single gene set

enrichit 0.0.6

• add vignette (2025-12-07, Sun)

enrichit 0.0.5

• implement multi-level GSEA algorithm (2025-12-06, Sat)

enrichit 0.0.4

• implement a simplified adaptive early-stopping GSEA algorithm (2025-12-05, Fri)
  • For each gene set:
  • 1. Run initial batch (e.g., 1000 permutations)
  • 2. If p-value > threshold (e.g., 0.05), stop
  • 3. If significant, increase permutations geometrically (2x, 4x, 8x...)
  • 4. Continue until p-value stabilizes or max permutations reached

enrichit 0.0.3

• implement ora_gson and gsea_gson (2025-12-05, Fri)
  • as replacement for enricher_internal and GSEA_internal
• mv helper functions and class definitions from DOSE to enrichit (2025-12-05, Fri)
  • to extend this package as the base package for the clusterProfiler family

enrichit 0.0.2

• gsea function (2025-12-04, Thu)
  • Gene Set Enrichment Analysis (GSEA) using C++ via Rcpp.

enrichit 0.0.1

• ora function (2025-12-03, Wed)
  • Fast Over-Representation Analysis (ORA) using C++ via Rcpp.