Key Deliverables & Statistical Frameworks
Cellular atlas embeddings, marker gene distributions, RNA velocity, pseudotime trajectory inference, ligand-receptor crosstalk, and spatial tissue architecture.
[DOMAIN 3: SINGLE-CELL & SPATIAL OMICS]
Plot 034 — Single-Cell UMAP Embedding
Use Case Application: Projects high-dimensional single-cell transcriptomes or ATAC profiles into 2D manifolds to resolve discrete cell types and states.
Required Input Data: Log-normalized, highly variable feature matrix (cells x HVGs) or batch-corrected PCA/Harmony latent coordinates.
Statistical & Algorithmic Method: Fuzzy simplicial set k-nearest-neighbor graph construction optimized via cross-entropy stochastic gradient descent (UMAP).
[DOMAIN 3: SINGLE-CELL & SPATIAL OMICS]
Plot 035 — Nebulosa Kernel Density Feature Plot
Use Case Application: Recovers spatial-like expression patterns of sparse or dropout-prone marker genes across single-cell UMAP embeddings.
Required Input Data: 2D UMAP/t-SNE cell coordinates paired with normalized single-cell gene expression vectors.
Statistical & Algorithmic Method: Weighted 2D Gaussian Kernel Density Estimation (KDE) smoothing expression intensity across local neighborhood manifolds.
[DOMAIN 3: SINGLE-CELL & SPATIAL OMICS]
Plot 036 — Single-Cell Marker Gene Dot Plot
Use Case Application: Concisely displays both detection prevalence (% expressing cells) and mean expression intensity across annotated cell clusters.
Required Input Data: Single-cell expression matrix and categorical cell-type cluster annotations for canonical marker genes.
Statistical & Algorithmic Method: Cluster-stratified percentage of non-zero cells (dot radius) and Z-scored mean log-normalized expression (color scale).
[DOMAIN 3: SINGLE-CELL & SPATIAL OMICS]
Plot 037 — Cluster Marker Violin Plot
Use Case Application: Compares full single-cell expression distributions, bimodality, and zero-inflation across cell subsets or treatment arms.
Required Input Data: Normalized single-cell gene expression vector stratified by cell cluster or donor condition.
Statistical & Algorithmic Method: Non-parametric Kernel Density Estimation mirrored symmetrically around group medians and interquartile ranges.
[DOMAIN 3: SINGLE-CELL & SPATIAL OMICS]
Plot 038 — Ridge (Joy) Density Plot
Use Case Application: Visualizes distributional shifts in pathway scores, module activities, or cytometric markers across ordered cell states.
Required Input Data: Continuous single-cell module score (AUCell / AddModuleScore) or surface protein (CITE-seq) intensity across groups.
Statistical & Algorithmic Method: Vertically offset Gaussian kernel density curves with quantile probability markers.
[DOMAIN 3: SINGLE-CELL & SPATIAL OMICS]
Plot 039 — RNA Velocity Streamline Vector Field
Use Case Application: Infers directional cell fate transitions and differentiation dynamics from unspliced vs. spliced mRNA kinetics.
Required Input Data: Single-cell spliced and unspliced count matrices (velocyto / kallisto-bustools) projected onto low-dimensional embeddings.
Statistical & Algorithmic Method: Dynamical or steady-state ODE modeling of transcription (alpha), splicing (beta), and degradation (gamma) rates (scVelo).
[DOMAIN 3: SINGLE-CELL & SPATIAL OMICS]
Plot 040 — PAGA Partition-Based Graph Abstraction
Use Case Application: Reconciles discrete cell clustering with continuous developmental topology by quantifying inter-cluster connectivity.
Required Input Data: Single-cell k-NN graph and discrete Leiden/Louvain cluster assignments.
Statistical & Algorithmic Method: Statistical connectivity measure comparing observed inter-cluster edges against random-graph null expectation.
[DOMAIN 3: SINGLE-CELL & SPATIAL OMICS]
Plot 041 — Pseudotime Differentiation Trajectory Curve
Use Case Application: Orders asynchronous cells along continuous lineage differentiation paths and identifies branching fate bifurcations.
Required Input Data: Low-dimensional diffusion map or PCA embedding with designated progenitor root cell state.
Statistical & Algorithmic Method: Principal curves / Minimum Spanning Tree smoothing (Slingshot / Monocle3) computing geodesic arc-length pseudotime.
[DOMAIN 3: SINGLE-CELL & SPATIAL OMICS]
Plot 042 — Pseudotime Gene Cascade Heatmap
Use Case Application: Reveals sequential waves of transcription factors and effector genes activated or repressed during lineage commitment.
Required Input Data: Single-cell expression matrix ordered by inferred pseudotime coordinates.
Statistical & Algorithmic Method: Generalized Additive Models (GAMs, tradeSeq) fitting smooth cubic splines of gene expression as a function of pseudotime.
[DOMAIN 3: SINGLE-CELL & SPATIAL OMICS]
Plot 043 — Alluvial / Sankey Cell-State Flow Diagram
Use Case Application: Tracks compositional shifts, cluster annotation concordance, or clonal transitions across longitudinal time points or methods.
Required Input Data: Contingency table of cell counts cross-tabulated across categorical states, time points, or clustering resolutions.
Statistical & Algorithmic Method: Proportional ribbon width scaling conserving total cell mass across stratified categorical axes.
[DOMAIN 3: SINGLE-CELL & SPATIAL OMICS]
Plot 044 — Ligand-Receptor Cell-Cell Communication Matrix
Use Case Application: Decodes paracrine and autocrine signaling networks between sender and receiver cell populations in the tissue microenvironment.
Required Input Data: Annotated scRNA-seq dataset paired with curated heteromeric ligand-receptor databases (CellPhoneDB / CellChat / LIANA).
Statistical & Algorithmic Method: Law of mass action interaction scoring with permutation-based null testing across shuffled cell cluster labels.
[DOMAIN 3: SINGLE-CELL & SPATIAL OMICS]
Plot 045 — Spatial Transcriptomics Tissue Overlay
Use Case Application: Maps gene expression programs, deconvolved cell types, or tumor-stroma niches directly onto histological H&E tissue sections.
Required Input Data: Spatial capture spot/cell coordinates (x, y), spatial count matrix (Visium / Xenium / CosMx), and H&E microscopy image.
Statistical & Algorithmic Method: Spatial spot deconvolution (RCTD / cell2location) and Moran's I spatial autocorrelation mapping.
[DOMAIN 3: SINGLE-CELL & SPATIAL OMICS]
Plot 046 — Spatial Neighborhood Enrichment Z-Score Heatmap
Use Case Application: Quantifies statistically significant co-localization or spatial avoidance between distinct cell phenotypes in intact tissue.
Required Input Data: Single-cell spatial coordinates and discrete cell phenotype annotations.
Statistical & Algorithmic Method: Delaunay triangulation / k-NN spatial graph permutation test (Squidpy) computing Z = (obs - mu_perm) / sigma_perm.
[DOMAIN 3: SINGLE-CELL & SPATIAL OMICS]
Plot 047 — Ripley's L(r) Spatial Clustering Function
Use Case Application: Evaluates multi-scale spatial clustering vs. dispersion of immune infiltrates or tumor foci across tissue radii.
Required Input Data: 2D point-pattern coordinates (x, y) of target cell populations within a defined tissue region of interest.
Statistical & Algorithmic Method: Variance-stabilized Ripley's K-function L(r) = sqrt(K(r)/pi) compared against Complete Spatial Randomness (CSR) Poisson envelope.
[DOMAIN 3: SINGLE-CELL & SPATIAL OMICS]
Plot 048 — Single-Cell Barcode Knee QC Plot
Use Case Application: Separates true intact cells from ambient RNA empty droplets during single-cell library preprocessing.
Required Input Data: Raw unfiltered cell-barcode by UMI count matrix sorted in descending order of total UMI counts.
Statistical & Algorithmic Method: Log-log rank-abundance spline inflection point detection (EmptyDrops / DropletUtils) modeling ambient RNA Dirichlet-multinomial profile.