Overview
The scSimEval Shiny Studio is a fully self-contained, browser-based graphical interface that gives researchers access to all 62 evaluation measures and every visualization in the package — without writing a single line of R code.
Launch it locally in R with:
library(scSimEval)
launch_scSimEval_app() # opens in your default browserFile size limit: The app accepts uploads up to 10 GB per file, which comfortably handles large-scale single-cell and multiomics datasets.
The studio is organized into seven primary navigation tabs in the top navigation bar, plus an eight-panel Visualizations module inside Tab 4. The sections below document every tab, every control, and all export options in the order they appear in the application.
Tab 1 — Home
The landing page introduces scSimEval Studio, explicitly highlighting simulation evaluation across single-cell RNA-seq (scRNA-seq), single-cell ATAC-seq (scATAC-seq), and single-cell multiomics (scRNA-seq + scATAC-seq) datasets against real biological references across 62 evaluation measures without requiring synthetic ground truth.
In the standalone web deployment (scSimEval_shiny), a
floating Google Analytics all-time pageview counter badge appears at the
bottom-right corner, continuously reporting container-lifetime visits
without UI latency.
The four KPI cards below the hero section display: 62 evaluation measures, 8 biological dimensions, 3 modalities (RNA / ATAC / Co-Assay), and the [0, 1] normalized fidelity scale.
Interactive operational mode cards showcase: 1. Pre-computed Benchmarks: 1-click loading of benchmark evaluations across multiple simulators and datasets. 2. Single-Cell RNA-seq Evaluation: Count matrix evaluation across statistical, clustering, and technical properties. 3. Multi-Omics (RNA + ATAC) Evaluation: Co-assay multiomics benchmarking assessing cross-modal coupling alongside unimodal properties. 4. Upload Pre-Computed Benchmark Table: Instant ingestion of previously calculated benchmark tables for reporting.
Tab 2 — Data Hub
Data ingestion and evaluation control centre. It features three comprehensive subtabs:
Subtab 1: Data Ingestion & Benchmark Studio
Choose one of four evaluation modes from the radio button panel on the left sidebar:
Mode 1 — Pre-Computed Benchmark (Demo Data)
Click Load Demo Benchmark to instantly populate the app with pre-computed multi-dataset results across Simulator 1 to Simulator 6 across all 62 measures. No file uploads required.
Package Example Datasets vs. Shiny App Demo Data: * R Package Example Datasets (
data/): Lightweight, runnable count matrices (example_scrna,example_scatac,example_multiomics, 60 features × 80 cells) designed for rapid command-line testing of individual evaluation functions in R console sessions. * Shiny App Demo Data (inst/shiny/scSimEvalApp/data/): A rich pre-computed benchmark results archive (6 simulators × 62 metrics × 2 datasets = 744 metric observations) accompanied by reference and simulated count matrices (100 genes × 50 cells). This gives web users an immediate, zero-latency tour of all tabs, ranking leaderboards, bubble matrices, and 8 visualization panels without requiring computationally intensive matrix evaluations.
Mode 2 — Single-Cell Evaluation (scRNA-seq / scATAC-seq)
Evaluate one or more unimodal count-matrix simulators across single
or multiple experimental datasets. * Reference Count
Matrix: Real empirical cells. Formats: .rds,
.csv, .tsv, .txt,
.xlsx. * Number of Single-Cell Simulators
(1–6): Generates matching upload cards with simulator name,
simulated count matrix, runtime (seconds), and peak RAM (MiB). *
Cell Type Labels (Optional): Vector / factor / 2-column
table (Cell_ID,Cell_Type). Unlocks +12–15 clustering and
biological signal metrics. * Batch Annotations
(Optional): Vector / factor / 2-column table
(Cell_ID,Batch). Unlocks +7 batch mixing metrics.
Mode 3 — Multiomics Evaluation (scRNA-seq + scATAC-seq)
- Paired Co-assay: Same individual cells in both modalities (10x Multiome, SHARE-seq). All 62 metrics including FOSCTTM, Match@1, and peak-to-gene linkage.
- Unpaired Profiling: Separate cells from the same tissue. Population-level cross-modal metrics evaluated; cell-pairing metrics safely skipped.
Metric Measurement Capacity by Upload Configuration
The Shiny app dynamically scales the number of quantitative metrics measured based on the uploaded files:
| Experiment Mode | Uploaded Inputs | Metrics Measured | Active Evaluation Categories & Scope |
|---|---|---|---|
| Single-Cell (scRNA / scATAC) | Count Matrix alone (no metadata) | ~23 Metrics | Cat I (Distributional, 14 metrics), Cat II (Correlations & Zeroes, 6 metrics), Cat VI (Trajectory, 2 metrics), Cat VIII (Scalability, 2 metrics). |
| Single-Cell (scRNA / scATAC) | Matrix + Cell Types | ~35–38 Metrics | Adds +12–15 metrics: Unlocks Cat III (Cellular Structure / ARI, NMI, AMI, Silhouette) and Cat V (Biological Signal / SimBench DE, DP, DV, top DEG Jaccard). |
| Single-Cell (scRNA / scATAC) | Matrix + Batch Info | ~30 Metrics | Adds +7 metrics: Unlocks Cat IV (Batch Mixing / Batch Silhouette, Shannon Entropy, PCR , CMS, ISI, Seurat mixing). |
| Single-Cell (scRNA / scATAC) | Full Metadata (Matrix + Cell Types + Batch) | Up to ~55 Metrics | Unlocks the complete unimodal suite across Categories I, II, III, IV, V, VI, and VIII. |
| Paired Multiomics (scRNA + scATAC) | RNA + ATAC Matrices alone | ~48 Metrics | Unimodal distributions for both modalities (~40 metrics) PLUS Cat VII (Cross-Modal Coupling: FOSCTTM, Match@1, generation, , peak-to-gene linkage). |
| Paired Multiomics (scRNA + scATAC) | Full Metadata (RNA + ATAC + Cell Types + Batch) | All 62 Metrics | Complete multiomics benchmark suite across all 8 canonical categories (Categories I–VIII). |
| Unpaired Multiomics (Independent cells) | RNA + ATAC Matrices alone | ~38–42 Metrics | Both unimodal distributions plus population coupling (co-expression module fidelity, Network Jaccard, cross-modality correlation). |
| Unpaired Multiomics (Independent cells) | RNA + ATAC + Cell Types (+ Batch) | Up to ~58 Metrics | Essential for unpaired: Unlocks cluster-matched label transfer, cell-type regulatory linkage, and technical donor integration. |
Mode 4 — Upload Pre-Computed Benchmark Table (.rds / .csv)
Load output previously saved from
evaluate_simulation_accuracy(),
evaluate_multiomics_accuracy(), or
evaluate_multiple_datasets().
Subtab 2: View Uploaded Data & Matrices
Interactive matrix viewer supporting browsing, filtering, and summary statistics across all loaded biological reference and simulated count matrices.
Subtab 3: Dataset Properties Summary & HPC Cluster Execution
- Dataset Properties Summary Table: Comprehensive per-matrix QC: Cells (N), Features (P), Sparsity (%), Cell Types, Batches, Median Library Size, Median Detected Features, Mean Expression.
- HPC Cluster Batch Script Generator: Export automated SLURM, PBS, or Bash cluster execution scripts to run heavy benchmarks offline on compute nodes.
Tab 3 — Comparative Bubble Matrix & Leaderboard
Subtab 1: Comparative Bubble Matrix Figure
The flagship visualization: every simulator × all 62 metrics in one colour-coded matrix. Bubble diameter is proportional to the normalized fidelity score. Scores ≥ 0.96 are rendered as bold square glyphs. Simulators are rank-ordered top-to-bottom by composite average score. * Multi-Dataset Selector Bar: Switch seamlessly between All Datasets (Cross-Dataset Average) or inspect individual datasets (Dataset 1, Dataset 2, …, Dataset N) separately. * Minimalist vs Publication Matrix: Toggle between a sleek, dot-matrix view and a fully annotated publication matrix. * Confirmation & Plot Action Button: Click Confirm & Generate Plot to render with app-wide loading indicators. * High-Resolution Exports & R Script: Download 600 DPI JPEG, vector PDF, or a Download R Script providing the complete analytical workflow with ggplot2 code.
Subtab 2: Fidelity Leaderboard
- Simulator Ranking Cards & Podiums: Visual podium ranking simulators by overall composite fidelity.
- Custom Category Weight Sliders: Interactively adjust the relative importance of each category to recalculate composite rankings according to specific biological priorities.
- Export Leaderboard: Download the tidy rankings table as CSV.
Tab 4 — Visualizations
Eight diagnostic sub-panels accessed via the pill navigation bar. Every sub-panel is equipped with the Global Multi-Dataset Selector Bar (allowing inspection of All Datasets Average or specific datasets), active dataset badges, Confirm & Generate Plot buttons, app-wide processing overlays, 600 DPI JPEG / PDF downloads, and a Download R Script workflow generator:
-
Sub-panel 1 — Evaluation Summary:
plot_evaluation_summary()— rank-ordered horizontal bar chart across all 8 categories and composite score. -
Sub-panel 2 — Distribution QC:
plot_distribution_qc()— grid comparing expression densities, ECDF curves, KS distances, library sizes, and zero-inflation curves between reference and simulated cells. -
Sub-panel 3 — Scalability Benchmark:
plot_scalability_benchmark()— runtime (seconds) and memory (peak RAM in MiB) dashboards with Pareto efficiency frontiers, throughput (cells/sec), and resource cost index. -
Sub-panel 4 — Metric Plots:
plot_metric_boxplots()/plot_category_metric_bars()/plot_individual_metric_bar()— score distribution boxplots and barplots across categories with Raw vs Normalized view toggles. -
Sub-panel 5 — Metric Heatmap:
plot_metric_heatmap()— method × metric grid with exact scores in bold text, direction-aware fill colours, and unsupervised hierarchical clustering. -
Sub-panel 6 — PCA Ordination:
plot_metric_pca()— simulators projected into performance space with discriminating metric vector loadings. -
Sub-panel 7 — MDS Metric Space:
plot_metric_mds()— Multi-Dimensional Scaling ordination mapping simulator algorithmic proximity. -
Sub-panel 8 — Cell Embeddings (UMAP, t-SNE &
PCA):
compute_dataset_embeddings()+plot_dataset_embeddings()— dimension reduction comparing reference cells directly to simulated cells, with cell-type cluster fidelity metrics.
Tab 5 — Benchmark Results & Export
The central hub for data exploration and multi-format exports:
Interactive Master Benchmark Scores Table
A full DT table with per-column dropdown search
boxes for instant filtering by Dataset, Method, and Category.
Explicitly displays: * Score_Raw: Raw unnormalized
metric values in physical and statistical units (e.g. KS distance,
Wasserstein distance, runtime in seconds, peak RAM in
MiB, Chi-Square values, Pearson correlation). *
Score: Direction-aware normalized score
where 1.0 represents top fidelity.
Tab 6 — Help & Getting Started
An in-app reference guide structured into seven sections: 1. Workflow Architecture & Pipeline — 4-step pipeline diagram and multi-dataset processing. 2. The Eight Evaluation Categories — tabular reference for all 62 metrics, highlighting peak RAM in MiB. 3. Metric Measurement Capacity by Data & Metadata Mode — detailed reference table documenting metric numbers and unlocked categories. 4. Score Normalization Pipeline — direction inversion, min-max scaling, and raw vs normalized score reporting. 5. Comprehensive Guide to App Tabs — detailed walkthrough of each application module. 6. Input Data Formats & Metadata Templates — matrix structures (genes × cells), cell-type and batch annotation vectors, and sample CSV downloads. 7. Computational Sizing, RAM (MiB/GiB), & Guidelines — hardware recommendations contrasting Cloud, Local Workstation, and HPC clusters.
Tab 7 — Contact
Information regarding the development team at Division of Agricultural Bioinformatics, ICAR-IASRI, maintainer contacts, GitHub issues link, and recommended citations.
R API Cross-Reference
Every Shiny control maps directly to a documented R function:
| Shiny Panel | R Function |
|---|---|
| Data Hub → Single-Cell | evaluate_simulation_accuracy() |
| Data Hub → Multiomics | evaluate_multiomics_accuracy() |
| Data Hub → Multiple Datasets | evaluate_multiple_datasets() |
| Bubble Matrix | plot_benchmark_bubble_matrix() |
| Evaluation Summary | plot_evaluation_summary() |
| Distribution QC | plot_distribution_qc() |
| Scalability Benchmark | plot_scalability_benchmark() |
| Metric Plots (all categories) | plot_metric_boxplots() |
| Metric Plots (by category) | plot_category_metric_bars() |
| Metric Plots (individual) | plot_individual_metric_bar() |
| Metric Heatmap | plot_metric_heatmap() |
| PCA Ordination | plot_metric_pca() |
| MDS Metric Space | plot_metric_mds() |
| Cell Embeddings |
compute_dataset_embeddings() +
plot_dataset_embeddings()
|
| Embedding Quality Metrics | compute_embedding_quality_metrics() |
For complete parameter descriptions see the Reference index.
#> R version 4.6.1 (2026-06-24 ucrt)
#> Platform: x86_64-w64-mingw32/x64
#> Running under: Windows 11 x64 (build 26200)
#>
#> Matrix products: default
#> LAPACK version 3.12.1
#>
#> locale:
#> [1] LC_COLLATE=English_India.utf8 LC_CTYPE=English_India.utf8
#> [3] LC_MONETARY=English_India.utf8 LC_NUMERIC=C
#> [5] LC_TIME=English_India.utf8
#>
#> time zone: Asia/Calcutta
#> tzcode source: internal
#>
#> attached base packages:
#> [1] stats graphics grDevices utils datasets methods base
#>
#> loaded via a namespace (and not attached):
#> [1] digest_0.6.39 desc_1.4.3 R6_2.6.1 fastmap_1.2.0
#> [5] xfun_0.60 cachem_1.1.0 knitr_1.52 htmltools_0.5.9
#> [9] rmarkdown_2.32 lifecycle_1.0.5 cli_3.6.6 sass_0.4.10
#> [13] pkgdown_2.2.1 textshaping_1.0.5 jquerylib_0.1.4 systemfonts_1.3.2
#> [17] compiler_4.6.1 tools_4.6.1 ragg_1.5.2 bslib_0.12.0
#> [21] evaluate_1.0.5 yaml_2.3.12 otel_0.2.0 jsonlite_2.0.0
#> [25] rlang_1.3.0 fs_2.1.0 htmlwidgets_1.6.4