yaara-dev/early-AD-data-code: early-AD-data-code 1.2.3

ExSeq Brain AD - Spatial Genomics Data Explorer & Analysis Scripts This repository contains an interactive web-based data explorer, Xenium HTML viewers, and computational analysis pipelines for spatial transcriptomics data from the ExSeq Brain AD study. 🚀 Data Explorer The full interactive data explorer is available via GitHub Pages: 👉 https://yaara-dev.github.io/early-AD-data-code/ Click the link above to explore the spatial genomics data with interactive 2D and 3D visualizations, filtering options, and comprehensive dashboards. 📋 For detailed information about data structure, column descriptions, and file relationships, see data/README.md Features Multi-sample support: Browse between different samples Interactive 2D scatter plot: Visualize spatial data with x/y coordinates Interactive 3D scatter plot: Explore data in three dimensions with rotation and zoom Comprehensive dashboard: View statistics, heatmaps, distributions, and more Filtering: Filter by region, gene, cell type, and Z-slice Statistics: View total records, unique genes, regions, and visible points Tooltips: Hover over points to see detailed information Samples Available WT: WT_1, WT_2.1, WT_2.2, WT_3, WT_4, WT_5, WT_6.1, WT_6.2 5xFAD: 5xFAD_1.1, 5xFAD_1.2, 5xFAD_2, 5xFAD_3, 5xFAD_4, 5xFAD_5, 5xFAD_6.1, 5xFAD_6.2 All samples use a separated/normalized CSV format with 5 files per sample in data/viewer_normalized/ (for the web viewer). 📊 Analysis Scripts This repository includes four computational analysis pipelines for spatial transcriptomics data: Cell Typing - Supervised STELLAR encoder label transfer for ExSeq cell-type annotation Expression, Spatial Distribution & Moran's I - MATLAB scripts for differential expression, localization/spatial distribution, and Moran's I RNA Velocity - Analysis of RNA velocity and spatial cell-state transitions SVG Neighborhood Analysis - Post-CELINA analysis of spatially variable genes and co-regulated modules Each analysis pipeline is self-contained with its own documentation, requirements, and example data. 📖 For detailed information about each analysis, see analysis_scripts/README.md Each analysis directory contains its own README with: Installation instructions Input data format requirements Step-by-step usage examples Output format descriptions 🔬 Xenium HTML viewers Self-contained Xenium Ranger analysis-summary reports for the Xenium samples. Open any .html file in a web browser (no server required) to inspect run quality, transcript and cell metrics, and spatial summaries. Files live in data/xenium_html/ in this repository. The same reports are also included in both Zenodo deposits (code and analysis data). WT: WT_5, WT_6_section1, WT_6_section2 5xFAD: 5xFAD_5_section1, 5xFAD_5_section2, 5xFAD_6 | File | |------| | WT_5_analysis_summary.html | | WT_6_section1_analysis_summary.html | | WT_6_section2_analysis_summary.html | | 5xFAD_5_section1_analysis_summary.html | | 5xFAD_5_section2_analysis_summary.html | | 5xFAD_6_analysis_summary.html | These reports are for visual exploration of the Xenium runs. They are not used by the analysis scripts. For Developers The following information is only relevant if you want to modify or extend the visualization code. Project Structure Data Explorer Components: index.html - Main visualization file (contains 2D view, 3D view, and dashboard) data/csvs/manifest.json - Sample manifest file (auto-generated) data/viewer_normalized/ - Separated normalized CSVs for the web viewer (5 files per sample) data/exseq/, data/xenium/, data/xenium_full_sections/ - Analysis tables (download from the data deposit; see data/README.md) data/xenium_html/ - Xenium Ranger HTML reports (in this repository, and also in both Zenodo deposits) data_explorer_scripts/ - Data generation and normalization scripts generate_manifest.py - Generates manifest.json from CSV files add_cell_types.py - Adds cell type information to CSV files server.py - Local development server with cache control Analysis Scripts: analysis_scripts/ - Computational analysis pipelines cell_typing/ - Supervised STELLAR encoder label transfer (Python) expression_spatial_distribution_morans_i/ - Expression, spatial distribution/localization, and Moran's I (MATLAB) rna_velocity/ - RNA velocity analysis (Python) svg_neighborhood_analysis/ - SVG neighborhood analysis (Python) Zenodo archives Code (this repository release, including Xenium Ranger HTML reports): https://doi.org/10.5281/zenodo.17568649 Analysis data (ExSeq / Xenium tables, and the same Xenium Ranger HTML reports): https://doi.org/10.5281/zenodo.21903277 Large ExSeq/Xenium tables are not stored in git. To run the MATLAB analyses: Clone this repository (or download the code Zenodo archive). Download the data deposit. Unpack so these folders are populated: data/exseq/ data/xenium/ data/xenium_full_sections/ Run the scripts under analysis_scripts/expression_spatial_distribution_morans_i/ — they already point at those paths. Xenium Ranger HTML reports are in data/xenium_html/ in this repository (open in a browser). They are also included in both Zenodo deposits. They are not used by the MATLAB scripts. Data Format The visualization uses a normalized CSV structure with 5 files per sample: *_points.csv - Point data (point_id, gene, Z, X, Y, fov) *_regions.csv - Region information (region_id, region_name, region_area, region_proportion) *_points_regions.csv - Point-to-region mapping (point_id, region_id) *_cells.csv - Cell information (cell_id, cell_type) *_points_cells.csv - Point-to-cell mapping (point_id, cell_id) 📋 For detailed information about data structure, column descriptions, and file relationships, see data/README.md Local Development Generate the manifest file (if data changes): python3 data_explorer_scripts/generate_manifest.py Run a local server: python3 data_explorer_scripts/server.py Or use Python's built-in server: python3 -m http.server 8000 Open in browser: Go to http://localhost:8000 The visualization will automatically load the manifest and first sample Deployment The project uses GitHub Actions for automatic deployment. The workflow (.github/workflows/deploy.yml) automatically: Generates the manifest file Deploys to GitHub Pages on changes to index.html, data/viewer_normalized/, or data_explorer_scripts/generate_manifest.py Requirements For Data Explorer: A modern web browser (Chrome, Firefox, Safari, Edge) Python 3 (for running scripts and local development) Separated normalized CSV files in data/viewer_normalized/ folder For Analysis Scripts: See individual analysis README files in analysis_scripts/ for specific requirements Most Python analyses require Python 3.8+ with scientific packages (pandas, numpy, scipy, etc.) Cell typing (cell_typing/) requires Python packages listed in analysis_scripts/cell_typing/environment.yml Expression / spatial distribution / Moran's I (expression_spatial_distribution_morans_i/) requires MATLAB (see that folder's README for toolboxes)