Customer Data Organization¶

This page describes how data is organized within Customer Workspaces on Latch.

Workspaces are organized by Workflow output: each top-level directory holds the results of one Workflow, and within it each project or run gets its own subdirectory. Directory names encode the modality and stage — for example epi_optimize_archr/ holds ATAC (epigenomic) optimization results produced with ArchR.

Workspace layout¶

<workspace>/
├── spatials/                        # tissue images + metadata (AtlasXBrowser)
│
├── fastq2frags/                     # ATAC: FASTQ → fragments
├── cram2frags/                      # ATAC: CRAM → fragments
│
├── epi_optimize_archr/             # ATAC: parameter sweep (ArchR)
├── epi_optimize_snap/              # ATAC: parameter sweep (SnapATAC2)
│
├── epi_analysis_archr/             # ATAC: secondary analysis (ArchR)
├── epi_analysis_snap/              # ATAC: secondary analysis (SnapATAC2)
├── compare_outs/                    # ATAC: differential comparisons
│
├── rna_preprocessing/               # RNA: alignment + QC
├── rna_analysis/                    # RNA: secondary analysis
│
└── copro_integration_analysis/      # Co-Profiling: epigenome × transcriptome

Directory reference¶

Directory	Produced by	Contents
`spatials/`	AtlasXBrowser	The Spatial folder per run — tissue images, `tissue_positions_list.csv`, and metadata.
`fastq2frags/`	ATX epigenomic preprocessing	Filtered FASTQs, Chromap alignment, the fragments file, and QC metrics/reports.
`cram2frags/`	cram2frags	Fragments produced from CRAM input (e.g. Ultima runs) rather than FASTQ.
`epi_optimize_archr/`	optimize archr	Parameter-sweep comparison figures and metrics (ArchR).
`epi_optimize_snap/`	optimize_snap	Parameter-sweep comparison figures and metrics (SnapATAC2).
`epi_analysis_archr/`	create ArchRProject	The ArchRProject, Seurat/AnnData objects, coverage tracks, peak BEDs, and analysis tables from ArchR.
`epi_analysis_snap/`	ATX_snap	Combined AnnData, gene-accessibility and motif results, and analysis tables from SnapATAC and ArchR.
`compare_outs/`	compare clusters	Differential gene, peak, and motif results between user-specified groups.
`rna_preprocessing/`	RNAQC	STARsolo alignment, the gene-expression matrix, MultiQC report, and contamination screen.
`rna_analysis/`	optimize_wt	Clustered AnnData, marker genes, spatially variable genes, and figures.
`copro_integration_analysis/`	atx_glue	SpatialGlue-integrated objects, coverage tracks, peak-to-gene links, correlation results.

Raw FASTQs are not delivered by default

Instead of raw sequencing reads, workspaces receive the filtered FASTQs returned by ATX epigenomic preprocessing — the reads that survive linker filtering, under fastq2frags/<run_id>/filtered_fastqs/.

Inside a Workflow directory¶

Each top-level directory contains one subdirectory per run (for run-level Workflows) or per project (for Workflows that merge multiple runs together). For example, epigenomic preprocessing writes one <run_id>/ folder per run:

fastq2frags/
└── <run_id>/
    ├── fragments.tsv.gz                # the fragments file
    ├── fragment_analysis_report.html   # QC report
    ├── <run_id>_cell.bw                # coverage track
    ├── filtered_fastqs/                # linker-filtered reads + bbduk stats
    ├── chromap_output/                 # alignment (BED, fragments, index, log)
    ├── run_metrics/                    # per-tixel metrics and QC plots
    └── pycistopic_metrics/             # summary statistics and peaks

See each Workflow's Outputs section for the full contents of its directory.

Old structure (superseded)

Earlier Customer Workspaces grouped data by processing stage rather than by Workflow. These folders are no longer used for new deliveries, but you may still encounter them in older workspaces:

Folder	Contents
`Raw_Data`	Raw FASTQ files, spatial folders, preprocessing outputs for ATAC and RNA.
`Processed_Data`	Outputs from the secondary processing pipeline.
`Optimized_Data`	Preprocessing results from shallow-sequenced optimization experiments (usually bulks).