Nodes by modality.

Load an .h5ad file into AnnData.

Load 10X Genomics HDF5 (.h5) from Cell Ranger output.

Load 10X Genomics MTX directory (matrix.mtx + features + barcodes).

Load a CSV/TSV expression matrix (cells × genes).

Load a Loom file. Use Load Velocyto Loom for RNA velocity layers.

Load a generic text expression matrix.

Load a velocyto .loom for RNA velocity (spliced/unspliced/ambiguous layers).

Load STARsolo Velocyto MTX directory (spliced/unspliced/ambiguous).

Filter cells/genes by QC metrics. Unified node (replaces qcMetrics + qcVisualization).

Detect doublets using Scrublet. Run early in the pipeline.

Normalize total counts per cell (scanpy normalize_total).

log(x+1) transformation. Run after Normalize, before HVG/PCA.

Identify highly variable genes (HVGs) that drive biological variation.

Scale to unit variance and zero-center. Required before PCA. Auto-saves adata.raw.

Principal Component Analysis — required before Neighbors/UMAP/Harmony.

Harmony batch correction — outputs to obsm["X_pca_harmony"]. Set use_rep="X_pca_harmony" in Neighbors after.

k-nearest neighbor graph on PCA space. Required for clustering and UMAP.

UMAP 2D embedding for visualization. Local neighborhood structure preserved; global distances are NOT meaningful.

t-SNE embedding (alternative to UMAP). Good for distinct clusters; slower than UMAP.

Leiden or Louvain community detection on the kNN graph, or K-Means with a fixed number of clusters on the embedding.

Rename or merge cluster labels for cell type annotation.

LLM-based cell type annotation (OpenAI / Anthropic / Gemini). No reference dataset needed.

Per-cell type annotation using CellTypist pretrained logistic-regression models — no LLM and no custom reference required.

Reference-based label transfer via kNN mapping on joint PCA.

Compare 2+ categorical annotation/cluster columns and visualise their agreement — a QC step to judge whether an automatic annotation can be trusted.

Differential expression — ranks all genes per cluster (Wilcoxon by default). Replaces "Find Markers" naming.

Extract top N DEG genes per cluster as a deduplicated gene list (renamed from "Extract Gene List").

GO/KEGG enrichment per cluster using gseapy + Enrichr REST API.

Bulk pathway enrichment (gseapy) — preranked GSEA or ORA.

Infer the activity of 14 signalling pathways from signed downstream expression footprints.

Score 50 MSigDB Hallmark biological programs from the expression matrix.

Infer transcription-factor activity from curated, signed TF-target regulons.

Cell-cell ligand-receptor interaction analysis. Modes: "run" (in-app) or "load" (read existing CLI output).

Consensus ligand-receptor scoring (rank_aggregate) combining CellPhoneDB, NATMI, Connectome, logFC and more into one ranking of interactions between cell groups.

Diffusion map embedding — captures continuous transitions. Required before DPT.

Partition-based Graph Abstraction — coarse-grained trajectory topology between clusters.

Diffusion Pseudotime — orders cells along a continuous trajectory (0 = root, 1 = differentiated).

Force-directed graph layout (ForceAtlas2 / Fruchterman-Reingold).

Estimate RNA velocity from spliced/unspliced counts (scvelo).

Compute velocity transition graph for stream/arrow plots.

Latent time from dynamical RNA velocity. More accurate than DPT when velocity available.

Identify terminal/initial states from velocity transition matrix.

Compute cell-cell transition probabilities (Markov chain). Kernel: cytotrace / pseudotime / velocity / connectivity.

GPCCA — identify terminal/initial macrostates and compute fate probabilities.

Identify genes driving differentiation toward each cell fate (correlation with fate probabilities).

UMAP/t-SNE/PCA scatter colored by metadata or gene expression.

Cluster composition, proportions, embedding overlay, dendrogram.

Heatmap, dotplot, violin, stacked violin, matrixplot, tracksplot, volcano of marker genes.

Highly variable gene dispersion plot.

PCA elbow (variance ratio), scatter, loadings, overview.

Histogram, scatter, embedding overlay of doublet detection results.

PAGA graph, DPT pseudotime, diffusion map, draw_graph plots.

Stream, arrow, phase portrait, speed, coherence, latent time.

Plot fate analysis results (fate probs, macrostates, terminal/initial states, circular projection, coarse transition, lineage drivers).

GAM-smoothed expression trends along differentiation lineages.

Enrichr GO/KEGG enrichment as barplot or dotplot.

Visualize CellPhoneDB results — heatmap (interaction counts), dotplot (LR pairs for specific cell types), or chord (network).

Dotplot of LIANA consensus ligand-receptor results — dot size = interaction specificity, colour = expression magnitude.

Concatenate multiple AnnData objects with per-input labels and live pre-merge preview.

Attach external metadata CSV/TSV/Excel to adata.obs with auto-match suggestions and live match-rate preview.

Filter cells or genes using multi-condition AND/OR builder with live preview.

Assign cells to custom categorical groups via multi-condition rules; same label across groups acts as OR.

Split cell IDs by delimiter into new obs columns (e.g. B_cac10_AAACCTGAG → type/patient/barcode).

Browse adata.obs and apply small obs edits (drop / rename / dtype convert) in one node. Dedicated UI: multi-select autocompletes, rename pair editor, inline obs preview table, per-column summary accordion (dtype, n_unique, n_missing, value_counts top N bar chart, sample uniques).

Save AnnData to .h5ad.

Export analysis results to CSV — 15 export types (markers, pathway, obs, var, lineage_drivers, fate_probabilities, cpdb, bulk_de, gsea, atac_*, dmp, dmr, transcripts, xenium).

Reads a whole 10x Visium SpaceRanger folder — counts, real coordinates, true scalefactors, and the H&E image.

Loads one chosen square-bin resolution from a Visium HD run, with a bin-count guard applied before anything is read into memory.

Reads a 10x Xenium cell-feature matrix with its aligned micron centroids, without pulling in whole-slide morphology images.

Loads Vizgen MERSCOPE cell-by-gene counts with their metadata, micron coordinates, and optional coordinate transformation.

Loads flat NanoString CosMx exports, using slide-wide global coordinates as the canonical spatial system.

Load single-molecule transcript locations (Xenium / MERFISH / seqFISH) into adata.uns["transcripts"].

Add spatial X/Y from CSV/TSV directly to obsm["spatial"]. Has its own loader — no Load CSV needed.

Load an .h5ad file into AnnData.

Load 10X Genomics HDF5 (.h5) from Cell Ranger output.

Load 10X Genomics MTX directory (matrix.mtx + features + barcodes).

Load a CSV/TSV expression matrix (cells × genes).

Filter low-density spatial regions (Xenium edges, bubbles, control probe noise) using a grid.

Extracts per-spot intensity, histogram, and texture descriptors from the tissue image already attached to the data.

Filter cells/genes by QC metrics. Unified node (replaces qcMetrics + qcVisualization).

Normalize total counts per cell (scanpy normalize_total).

log(x+1) transformation. Run after Normalize, before HVG/PCA.

Identify highly variable genes (HVGs) that drive biological variation.

Scale to unit variance and zero-center. Required before PCA. Auto-saves adata.raw.

Principal Component Analysis — required before Neighbors/UMAP/Harmony.

Harmony batch correction — outputs to obsm["X_pca_harmony"]. Set use_rep="X_pca_harmony" in Neighbors after.

k-nearest neighbor graph on PCA space. Required for clustering and UMAP.

UMAP 2D embedding for visualization. Local neighborhood structure preserved; global distances are NOT meaningful.

Leiden or Louvain community detection on the kNN graph, or K-Means with a fixed number of clusters on the embedding.

Rename or merge cluster labels for cell type annotation.

LLM-based cell type annotation (OpenAI / Anthropic / Gemini). No reference dataset needed.

Per-cell type annotation using CellTypist pretrained logistic-regression models — no LLM and no custom reference required.

Reference-based label transfer via kNN mapping on joint PCA.

Read-only interactive Plotly WebGL scatter. Color by obs column or gene, select with lasso, box or polygon, highlight groups, inspect stats — all without modifying adata.

Lasso/box-select tissue regions and label them as obs categorical column. Replaces spatialInteractiveViewer.

Spatial neighbor graph based on physical coordinates. Required for all spatial statistics.

Test if cell types co-localize or avoid each other (squidpy nhood_enrichment).

Reports the observed contact matrix between clusters plus each cluster's network role on the spatial graph.

Spatial co-occurrence at multiple distances.

L-R interaction analysis using squidpy ligrec (CellPhoneDB-like permutations).

Scores ligand-receptor pairs per spot against the spatial graph, mapping where signalling is co-located.

Segments tissue into contiguous domains by blending each cell's expression with its physical neighbours.

Groups cells by the mix of cell types surrounding them, revealing recurring tissue microenvironments.

Estimates the cell-type mixture inside each spot with a transparent non-negative least-squares fit against a labelled reference.

Moran's I, Geary's C, or sepal diffusion time — identifies spatially variable genes (SVGs).

Finds spatially autocorrelated genes and groups them into coordinated local-expression modules using Hotspot.

Ripley's L/F/G statistics — measures spatial clustering patterns.

Spatial scatter and analysis overlays (nhood, co-occurrence, ripley, autocorr, ligrec).

UMAP/t-SNE/PCA scatter colored by metadata or gene expression.

Attach external metadata CSV/TSV/Excel to adata.obs with auto-match suggestions and live match-rate preview.

Filter cells or genes using multi-condition AND/OR builder with live preview.

Assign cells to custom categorical groups via multi-condition rules; same label across groups acts as OR.

Split cell IDs by delimiter into new obs columns (e.g. B_cac10_AAACCTGAG → type/patient/barcode).

Browse adata.obs and apply small obs edits (drop / rename / dtype convert) in one node. Dedicated UI: multi-select autocompletes, rename pair editor, inline obs preview table, per-column summary accordion (dtype, n_unique, n_missing, value_counts top N bar chart, sample uniques).

Save AnnData to .h5ad.

Export analysis results to CSV — 15 export types (markers, pathway, obs, var, lineage_drivers, fate_probabilities, cpdb, bulk_de, gsea, atac_*, dmp, dmr, transcripts, xenium).

Load a generic bulk count matrix (CSV/TSV/H5AD). Genes × samples typically — enable Transpose.

featureCounts .txt or STAR ReadsPerGene.out.tab. Auto-handles header lines and STAR strand selection.

Load Kallisto abundance.tsv or Salmon quant.sf. Format auto-detected. Single-file loader; merge per-sample with Merge Data.

Load an .h5ad file into AnnData.

Attach external metadata CSV/TSV/Excel to adata.obs with auto-match suggestions and live match-rate preview.

Assign cells to custom categorical groups via multi-condition rules; same label across groups acts as OR.

Sample-level QC — library size, gene count, IQR-based outlier detection.

Remove lowly expressed genes (e.g. CPM > 1 in ≥2 samples).

CPM / TMM / DESeq2 size-factor / VST. Raw counts always saved to layers["counts"].

PyDESeq2 differential expression between two conditions (t-test fallback).

Bulk pathway enrichment (gseapy) — preranked GSEA or ORA.

Infer the activity of 14 signalling pathways from signed downstream expression footprints.

Score 50 MSigDB Hallmark biological programs from the expression matrix.

Infer transcription-factor activity from curated, signed TF-target regulons.

Volcano, MA, DEG heatmap, sample distance, library size.

Filter cells or genes using multi-condition AND/OR builder with live preview.

Browse adata.obs and apply small obs edits (drop / rename / dtype convert) in one node. Dedicated UI: multi-select autocompletes, rename pair editor, inline obs preview table, per-column summary accordion (dtype, n_unique, n_missing, value_counts top N bar chart, sample uniques).

Save AnnData to .h5ad.

Export analysis results to CSV — 15 export types (markers, pathway, obs, var, lineage_drivers, fate_probabilities, cpdb, bulk_de, gsea, atac_*, dmp, dmr, transcripts, xenium).

Load Illumina methylation IDAT (450K / EPIC v1 / EPIC v2 / Mouse). EPIC v2 auto-routes to mepylome (methylprep ceiling).

Load pre-processed beta-value matrix (CSV/TSV/TXT). Used when raw IDAT is unavailable.

Load an .h5ad file into AnnData.

Attach external metadata CSV/TSV/Excel to adata.obs with auto-match suggestions and live match-rate preview.

Assign cells to custom categorical groups via multi-condition rules; same label across groups acts as OR.

Per-sample / per-probe call rates with a built-in QC plot. "label" or "filter" mode.

Drop SNP-associated, sex chromosome, cross-reactive, low call rate probes. Mandatory before DMP.

NaN-aware between-array quantile normalization in beta space, or no-op. Raw-signal NOOB/SWAN preprocessing belongs in Load IDAT.

Differentially Methylated Position — vectorized Welch's t-test or limma-style moderated t-test.

Multi-group (3+) differential methylation — omnibus moderated F-test (empirical-Bayes shrinkage, limma-style) plus a reference contrast per group. Pure numpy, not R limma.

Differentially Methylated Regions (Bumphunter-style) — groups adjacent same-direction DMPs within max_gap_bp.

Fisher enrichment of significant DMPs against local BED or cached ENCODE SCREEN cCRE classes, with BH FDR and a 95% CI forest plot.

CpG-bias-corrected gene-set enrichment for DMP results (gometh-style two-step).

Genomic-coordinate Manhattan, capped/raw QQ and volcano, top/random/most-variable heatmap up to 10,000 probes, and direction-aware annotation distributions plus existing methylation plots.

Genome-browser-style locus plot for one DMR — per-sample beta + group means, DMR span shaded.

Filter cells or genes using multi-condition AND/OR builder with live preview.

Browse adata.obs and apply small obs edits (drop / rename / dtype convert) in one node. Dedicated UI: multi-select autocompletes, rename pair editor, inline obs preview table, per-column summary accordion (dtype, n_unique, n_missing, value_counts top N bar chart, sample uniques).

Save AnnData to .h5ad.

Export analysis results to CSV — 15 export types (markers, pathway, obs, var, lineage_drivers, fate_probabilities, cpdb, bulk_de, gsea, atac_*, dmp, dmr, transcripts, xenium).

Build peak×cell matrix from 10X fragments.tsv.gz. Calls peaks if no BED provided.

Load pre-computed peak×cell matrix (CSV or H5AD).

Cell Ranger ATAC MEX directory (matrix.mtx + peaks.bed + barcodes.tsv).

Load an .h5ad file into AnnData.

Per-cell QC — total fragments, peaks, FRIP score.

Call peaks with MACS3 (or top-N accessible regions fallback).

Latent Semantic Indexing (TF-IDF + randomized SVD). Standard ATAC dim-reduction.

k-nearest neighbor graph on PCA space. Required for clustering and UMAP.

UMAP 2D embedding for visualization. Local neighborhood structure preserved; global distances are NOT meaningful.

Leiden or Louvain community detection on the kNN graph, or K-Means with a fixed number of clusters on the embedding.

Differential chromatin accessibility between two cell groups (t-test + FDR).

TF motif enrichment in differentially accessible peaks via JASPAR / pyjaspar.

FRIP bar, fragment size, peak accessibility heatmap, ATAC UMAP, differential volcano.

Filter cells or genes using multi-condition AND/OR builder with live preview.

Browse adata.obs and apply small obs edits (drop / rename / dtype convert) in one node. Dedicated UI: multi-select autocompletes, rename pair editor, inline obs preview table, per-column summary accordion (dtype, n_unique, n_missing, value_counts top N bar chart, sample uniques).

Save AnnData to .h5ad.

Export analysis results to CSV — 15 export types (markers, pathway, obs, var, lineage_drivers, fate_probabilities, cpdb, bulk_de, gsea, atac_*, dmp, dmr, transcripts, xenium).