Epigenomics Plots

At a glance

Repository: atlasxomics/plots_epigenome · Display name: Epigenomics Plots · Modality: Epigenomics · Stage: Plots

Overview

Epigenomics Plots is an interactive Latch Plots notebook for exploratory analysis and figure generation from spatial epigenomic DBiT-seq experiments. Each notebook runs on its own cloud VM and is driven through a GUI in console.latch.bio; the interface is organized into tabs (each tab is a module), and you move between tabs to explore results.

How Plots connects to secondary analysis

Plots reads the objects produced by the epigenomic secondary analysis Workflows — ATX_snap or create ArchRProject. Specifically, it consumes:

• <project>_sm_ge.h5ad — the reduced combined gene-accessibility AnnData;
• <project>_sm_motifs.h5ad — the reduced combined motif-deviation AnnData;
• the coverage bigWig folders (cluster_coverages/, sample_coverages/, …).

These are exactly the *_sm_*.h5ad objects and *_coverages/ folders listed in those Workflows' outputs. Both Workflows also emit a Launch_Plots/artifact.json that opens their result directly in this Plots template.

Starting Plots

There are two ways to open a result in Plots:

1. Launch from a Workflow output (recommended). In Latch Data, open the secondary-analysis output folder (e.g. epi_analysis_snap/<project>/ or epi_analysis_archr/<project>/) and use its Launch_Plots/artifact.json / Open in Plots action. This starts a Plots notebook pre-loaded with that experiment's data.
2. Open Plots and select data manually. Start a new notebook from the Plots module in the Latch Console, then use the Select Data tab to point it at the *_sm_ge.h5ad and *_sm_motifs.h5ad files (plus the coverage folders). Setting a data path (re)initializes every other tab.

Full walkthrough

For a click-through walkthrough of launching and using Plots, see the AtlasXPlots Tutorial. The two methods above summarize the app's data-loading code; defer to the tutorial for exact click-steps.

Plotting modules

Each module is a tab in the Plots GUI.

Module	What it does
Welcome	Landing tab and report intro, plus the H5 Viewer for browsing the loaded AnnData — spatial and UMAP views, toggling gene vs. motif data, layout controls, and lasso-selecting cells into custom annotations (used to feed the Compare workflow).
Select Data	Points the notebook at the experiment's *_sm_ge.h5ad / *_sm_motifs.h5ad objects and coverage folders; specifying the path (re)initializes all other tabs.
Heatmap	Gene-accessibility and motif-enrichment heatmaps across clusters (from the genes_per_*_hm / motif_per_*_hm matrices).
Violin Plots	Distribution of a feature (gene accessibility or motif deviation) across clusters or groups.
Volcano	Volcano plots of differential gene accessibility / motif enrichment between groups.
Proportion Plot	Cluster / condition composition — the proportion of cells per cluster across samples or conditions.
Track Browser	Genome-browser view of coverage bigWig tracks, grouped by cluster, sample, or condition.
Motif Logo	Sequence-logo display of a motif's position-weight matrix (genome-specific seqlogo_* data).
Neighborhood Analysis	Spatial neighborhood / niche analysis — which clusters spatially co-localize.
Compare	Runs the compare clusters Workflow from within Plots: select two cell groups (via lasso / annotation), launch the Workflow, then fetch and visualize the gene, motif, and track-browser results.
Bulk Score Genes	Gene-set scoring and cell-type assignment — assign marker gene sets, score cells, review in the H5 viewer, and build a score heatmap.

Other modalities

The transcriptome and co-profiling paths have their own Plots apps — Transcriptome Plots and Co-Profiling Plots.