AtlasXomics Data Analysis Package

Process raw data locally, on our cloud-based platform, or with a combination of the two.

Download our analysis scripts and tutorials and process the data in your HPC or local environment

Screenshot of AtlasXomics GitHub repositories page showing eight repositories including spatial-atacseq_latch, downsample_latch, archr_latch, AtlasXbrowser, AtlasWebService, and AtlasWeb, with details on programming languages and popularity.

Upload your raw data to AtlasXomics’ cloud-based environment for automated analysis

Screenshot of a web-based interface for workflow management by AtlasXomics, Inc., showing a list of workflows including preprocessing ATAC-seq, optimize archr, clean, and create ArchRProject, with their descriptions and icons.

Raw data generated by the AtlasXomics platform

DBiT-seq combines microfluidics and Next-Generation Sequencing (NGS) to generate high-quality spatial data from a wide range of tissues. The assay produces microscopy and sequencing data for each sample.

Diagram describing the process of applying microfluidic chips to tissue samples. It shows tissue tissue sample, sequential overlay of microfluidic chips with barcode channels, a cross-section of the chip with PDMS channels, post-fluidic tissue image with fluorescent outlines, and a 3D illustration of tissue barcoding at each XY intersection, with steps numbered from 1 to 5.

  • The workflow starts by importing microscope images into AtlasXBrowser, our custom app for DBiT-seq image processing. This is done to identify on-tissue and off-tissue areas, and to produce the tissue position file which locates the spatial barcodes on tissue. Following this, the raw sequencing data undergoes processing through our custom epigenomic preprocessing and alignment workflow. This results in the generation of fragment files organized by pixel barcode. Once processed, the data can be inputted into ArchR or Signac, and then visualized in Seurat. Alternatively, the data can be further refined using our automated workflows available in our cloud-based workspace. Within this cloud environment, samples are combined and filtered. Cluster optimization is performed. Selected parameters are subsequently used to analyze all samples in a project, ultimately yielding differential epigenomic elements organized by clusters and samples. This data can be downloaded as processed Seurat and ArchR projects or can be viewed in our tailored Shiny app.

Flowchart illustrating a data processing pipeline with steps for image data and sequencing data analysis, involving cloud and local processing, optimization, clustering, and visualization.

Step 1: Generate position file with AtlasXBrowser 

Load post-microfluidics images into AtlasXBrowser either locally or in the Latch workspace. Identify on/off tissue pixels with a combination of automatic calling and manual selection.

Threshold image for automatic on/off tissue calling

A digital interface with a grayscale, pixelated photo of a person's face on the right and controls on the left, including sliders and buttons for image adjustments and analysis.

Manually select on tissue regions (red boxes)

A satellite image showing the surface of a moon or planet, overlaid with a grid of red squares used for analysis or mapping purposes.

Step 2: Process sequencing data and align to genome

FASTQ files can be processed on our cloud platform or locally with our Snakemake workflow. The process involves the following: reads are filtered on the correctness of ligation linker sequences; alignment and preprocessing is performed with Chromap; BED output files are converted into fragment.tsv.gz files for downstream analysis. 

Educational diagram showing workflow for uploading FASTQ files and running preprocessing using Snakemake, with screenshots of BaseSpace upload interface on the left and workflow execution on the right.

Transfer or download processed raw data for further analysis in R

Processed raw data (fragment and position files) can be downloaded and analyzed locally.  Data is compatible with many open-source packages including ArchR and Signac to identify clusters and differential epigenetic elements. Calculated clusters, motifs, and gene activity scores can be loaded into Seurat with the position files from Step 1 to visualize spatially.

See the ATX_epigenomics GitHub for a downstream processing tutorial.

Alternatively, the data can be processed in our cloud platform (Step 3).

Screenshots of an R programming interface displaying three colorful scatter plots labeled D01208, D01209, and D01210, with a legend below indicating multiple cluster colors.

Step 3: Select best clustering parameters and clean fragment file

Using the Optimize ArchR workflow, a wide range of clustering parameters can be visualized to determine the optimal set. In addition, using the Cleaning workflow, row and column artifacts can be minimized or removed prior to analysis

null

Step 4: Generate spatial epigenome data with selected parameters

Using one of our workflows, differential genes, peaks, and motifs by condition​, cluster, and sample are outputted with fully processed ArchR and Seurat objects. Data can be downloaded or loaded into our custom Shiny app for visualization.

Screenshot of a project creation interface for ArchProject, showing parameters, graph, executions, and development tabs. Files and datasets related to genomic research are listed on the right, including project files, CSV data, and text files.

Step 5: Visualize processed data in our interactive shinyApp

Load processed data into our interactive browser to easily explore spatially differentiated epigenetic elements in tissue.

Screenshots of two gene accessibility visualization tools comparing gene expression in two different datasets, with dropdown menus for selecting gene IDs and coordinates, color-coded scatter plots, and options to download the images.