Package index • msuiter

The object model (S7)

The central objects and the construction entry points (developer preview, the API marked experimental). The full function-to-concept-to-pillar map is the MSU-map article (articles/msu-map.html).

ms_variants(): Read variant calls into an MsVariants object
ms_catalog(): Construct an MsCatalog object
ms_signature(): Construct an MsSignature object
ms_refdb(): Construct an MsRefDb object
MsVariants(): MsVariants: validated variant calls with somaticness provenance
MsCatalog(): MsCatalog: channel-count matrix with a channel-registry snapshot
MsSignature(): MsSignature: extracted signatures with a catalog snapshot summary
MsFit(): MsFit: reference-fitted exposures with a reference snapshot summary
MsRefDb(): MsRefDb: versioned reference signature database
MsBenchmark(): MsBenchmark: benchmark grid results

Fitting and benchmarking

Reference-based fitting against a signature database, and the benchmark matrix that compares engines on a shared grid.

ms_fit(): Fit signature exposures against given signatures
ms_benchmark(): Construct an MsBenchmark object
fit_engine(): Run an engine through its registered implementation

The engine registry (the extension interface)

Third-party packages register engines via .onLoad(); the trust boundary and the onboarding checklist are in CONTRIBUTING.md.

register_ms_engine(): Register an engine in the msuiter engine registry
match_ms_engine(): Resolve string sugar to an MsEngine spec
required_pkgs(): Report the packages required by an engine
ms_engines(): List registered engines
MsEngine(): MsEngine: abstract engine specification

Diagnostics

The environment and provenance status report.

ms_sitrep(): Situation report for the msuiter installation

Engine implementations

The fitting entry points of the NMF/ARD/sparse trio of engines (dispatched through the engine registry).

ms_ard(): ms_ard(): ARD-NMF engine factory
MsArd(): MsArd: ARD engine specification
ms_nmf(): Create an NMF engine specification
MsNmf(): MsNmf: NMF engine specification
ms_sparse(): ms_sparse(): penalized sparse NMF engine factory
MsSparse(): MsSparse: penalized sparse NMF engine specification

Extraction and k selection

The tally/extract path and the k-selection evidence (CV/stability/reconstruction panel).

ms_tally(): Tally variants into a channel catalog
ms_extract(): Extract signatures from a catalog
ms_select_k(): Select the number of signatures from a rank grid
plot_k_cv_curve(): Plot the CV-error curve from K-selection evidence
plot_k_stability(): Plot the consensus-stability curves from K-selection evidence
plot_reconstruction_panel(): Original vs reconstruction with a per-channel residual strip

Monte-Carlo calibration

The calibration grid, the real-data sanity and the verdicts (incl. the coverage curves).

ms_calibration_grid(): Monte Carlo calibration grid for exposure confidence intervals
ms_calibration_sanity(): Real-data downsampling sanity harness for exposure confidence intervals
print(<ms_calibration_sanity>): Print a calibration sanity result
ms_calibration_verdict(): Structured window verdict over a calibration grid
plot_calibration_curve(): Plot the calibration coverage curve of a grid measurement

Bootstrap and stability

The bootstrap confidence intervals and the stability-evidence figures.

ms_fit_bootstrap(): Bootstrap percentile confidence intervals for signature exposures
plot_boot_ci(): Bootstrap CI error bars for one sample's exposures
plot_boot_distribution(): Bootstrap exposure distributions for one sample
plot_boot_stability(): Bootstrap stability bar face

Stratification, clustering and embedding

The hypermutator stratification, signature clustering and the exposure-space embedding.

ms_stratify_hypermutants(): Stratify hypermutated samples with the SigProfiler GMM cutoff
MsStratification(): MsStratification: hypermutant stratification decision
plot_hypermutator_burden(): Hypermutator burden histogram with the stored cutoff line
ms_cluster_signatures(): Cluster signatures by cosine similarity
ms_embed(): Embed samples by their exposure profiles
plot_embedding(): Scatter the sample embedding

Similarity and exposure comparison

The calibrated-similarity library, the comparison container, the exposure statistical test and their figure types.

ms_compare(): Calibrated comparison of two signature dictionaries
ms_comparison(): Construct an MsComparison object
MsComparison(): MsComparison: signature comparison results
ms_exposure_test(): Test per-signature exposure differences between sample groups
plot_similarity_heatmap(): Similarity heatmap from an ms_compare result
plot_exposure_heatmap(): Exposure heatmap (samples x signatures)
plot_exposure_riverplot(): Exposure riverplot (samples x signature flow)
plot_exposure_stacked(): Stacked exposure bars per sample

The visualization panel

The catalog profile, the COSMIC scatter, the reference comparison and the signature-catalog figures.

plot_catalog_profile(): Plot one profile column in COSMIC style
plot_cosmic_scatter(): Estimate vs COSMIC reference scatter
plot_reference_comparison(): Compare an estimated profile against a reference profile
plot_signature_catalog(): Plot a signature set as a facet grid

Import and export

The COSMIC/SigProfiler interoperability, the format conversion and the bundled reference-library faces.

ms_import(): Import a file into a catalog or signature set
ms_export(): Export a catalog or signature set in a file format
ms_convert(): Convert signatures between mutational opportunity representations
plot_opportunity_distribution(): Trinucleotide opportunity landscape
ms_refdb_bundled(): The bundled reference signature database (COSMIC v3.6)
ms_refdb_manifest(): The governance manifest of the bundled reference database
ms_catalog_registry(): The committed catalog-currency digest registry
ms_catalog_currency(): Verify the pinned COSMIC catalog snapshot for currency
ms_catalog_currency_update(): Refresh the catalog-currency cache from the official source documents
print(<ms_catalog_currency_report>): Print a catalog-currency report

Input construction and utilities

The variant/breakpoint input constructors, downsampling, simulation and the presence test.

ms_segments(): Read allele-specific copy-number segments into a typed table
ms_sv(): Read structural-variant calls (BEDPE) into a typed table
ms_downsample(): Downsample a catalog to a common depth
ms_simulate(): Simulate a catalog from a signature dictionary
ms_test_presence(): Cohort presence test for signature exposures (BH-corrected LRT)

QC and reporting

The QC report, its sibling sentries surfaces and the HRD report faces.

ms_qc_report(): Per-sample QC report
ms_qc_burden(): Per-sample burden distribution table
print(<MsQcBurden>): Print a burden distribution table
ms_qc_sentries(): Per-sample gt/ct mechanism sentries
print(<MsQcSentries>): Print a gt/ct sentries table
ms_qc_mechanism_channels(): The SBS96 mechanism-channel roster
print(<MsQcMechanismChannels>): Print a mechanism-channel roster
ms_qc_artifact_roster(): The cataloged COSMIC artifact roster
print(<MsQcArtifactRoster>): Print an artifact roster
ms_qc_roster_share(): Per-sample artifact-roster exposure share
print(<MsQcRosterShare>): Print a roster exposure-share table
ms_qc_provenance(): The QC provenance checklist
print(<MsQcProvenance>): Print a provenance checklist
plot_qc_dashboard(): QC dashboard: burden bars plus the sentry shares
plot_pseudosig_contribution(): Artifact-signature contribution ranking
ms_hrd_report(): HRD report from allele-specific segment calls
print(<MsHrdReport>): Print an HRD report
as.data.frame(<MsHrdReport>): Coerce an HRD report to its per-sample table
ms_treatment_report(): Treatment-forensics report: bootstrap-CI association plus clonal timing
print(<MsTreatmentReport>): Print a treatment report
as.data.frame(<MsTreatmentReport>): Coerce a treatment report to its association table

The benchmark-matrix driver

The scenario constructor and the driving loop (the naming ruling is in the 2026-10-10 devlog record, articles/devlog-2026-10-10-m7-benchmark-ergonomics.html).

ms_benchmark_grid(): Construct one benchmark scenario
ms_run_benchmark(): Run the benchmark grid