README
======

Dataset title:
Smart-Seq3Xpress scRNA-seq data from human skin across 4 donors and 4 body sites

Recommended Citation:  
Yuan, H., Annusver, K., & Kasper, M. (2026). Smart-seq3xpress scRNA-seq data from human skin across 4 donors and 4 body sites (Version 1) [Data set]. Karolinska Institutet. https://doi.org/10.48723/y92n-w439
See the Researchdata.se catalogue entry for this dataset - accessible via the DOI link above - for compete metadata, including the dataset description, contact information and information about how to obtain the dataset. 

Responsible organization:
Karolinska Institutet

Description: Smart-Seq3Xpress scRNA-seq datasets from 4 donors, covering 4 bodysites each. Healthy human skin biopsies were obtained from four adult donors, including two female and two male donors aged 30–45 years. From each donor, biopsies were collected from four anatomical sites: upper back, occipital scalp, plantar skin and inner upper arm. Skin samples were collected using a 4 mm punch biopsy. Next, tissue was dissociated with Miltenyi Whole Skin Dissociation Kit and these isolated cells were sorted (only live cell) into separate wells in 384-well lysis plates (28 or 32 plates per sample) for Smart-Seq3Xpress library preparation and subsequent sequencing. Prepared libraries were sequenced using BGI DNBSEQ-T7 and BGI DNBSEQ-G400 platforms. Sequencing reads were aligned to the human GRCh38 reference genome and processed using zUMIs.

Access restrictions: This dataset contains human genomic data and is subject to legal protections. This dataset can be requested from the Researchdata.se research data catalog using the link above.

Essential information for dataset reuse.
----------------------------------------
This dataset contains Smart-Seq3Xpress single-cell RNA-seq data from healthy adult human skin. The data are organized by donor and anatomical body site. Each donor contributed samples from multiple body sites, and donor identity, sex, age and anatomical site should be considered as relevant metadata variables when reusing the dataset.

The dataset is intended for reuse in analyses of human skin cell-type composition, transcriptional heterogeneity, donor-to-donor variation and body-site-associated gene expression differences.

Single cells were isolated by fluorescence-activated cell sorting and processed using Smart-Seq3Xpress library preparation. Sequencing was performed on BGI DNBSEQ-T7 and DNBSEQ-G400 platforms. Reads were aligned to the human GRCh38 reference genome and processed using zUMIs (version 2.9.7).

Reuse of the raw sequencing data requires software compatible with Smart-seq-style single-cell RNA-seq processing. Processed data can be analysed using standard single-cell RNA-seq analysis frameworks (e.g., Scanpy, Seurat or equivalent tools).


Datafile descriptions
----------------------
FASTQ files (`*.fq.gz`):

Gzip-compressed FASTQ files contain raw sequencing reads and their corresponding per-base quality scores. Files are organized by read type.

* `read1` / `_1.fq.gz`: contains Read 1 sequences. In this dataset, Read 1 files are concatenated without sequence modification.

Each FASTQ record consists of four lines: read identifier, nucleotide sequence, separator line, and base quality string. 

Files starting with "V" come from G400 sequencer, Files starting with "E100099999" come from T7 sequencer.


zUMIs configuration file (`*.yaml`):
The YAML file contains the run configuration used for processing the sequencing data with zUMIs. It defines input FASTQ files, output locations, barcode and UMI handling, read layout, genome reference files, STAR alignment settings, and counting parameters.
This file can be used to reproduce the original zUMIs preprocessing run, provided that the required FASTQ files and reference genome resources are available.


References for mapping:
Files in the References subfolder are used for mapping reads onto genome annotations. The files include a directory with the used STAR index and a separate annotation .gtf file.


Other files (`*.txt`, `*.csv`):
* `fastq_info_processing_for_zUMIs_run.txt`: script used to pre-process (merge) fastq files from G400 sequencer for a subsequent mapping with zUMIs
* `bc_whitelist.txt`: barcode whitelist file used by zUMIs
* `bc_annotation_full.txt`: file mapping donor and bodysite information to specific barcodes (cells)
	- `plateID`: Smart-seq3xpress plate identifier.
	- `indices`: Library/index set used for the plate.
	- `XC_DNBPE`: Cell barcode sequence used to identify the well/cell in DNBSEQ paired-end data.
	- `WellID`: Well position within the Smart-seq3 plate.
	- `donor`: Donor identifier.
	- `group`: Group of plates belonging to the same donor and bodysite combination.
	- `bodysite`: Anatomical sampling site for the cells in the corresponding plate/well.

* `Smart-seq3xpress_donor_metadata.csv`: donor specific metadata

Reference information
---------------------
Abbreviations:
The following abbreviations are used in file names, metadata files, or processing configuration files:

FASTQ / .fq.gz: compressed raw sequencing read files.
read 1: first sequencing read.
read 2: second sequencing read.
zUMIs: processing pipeline used for alignment, barcode/UMI handling, and generation of count matrices from the raw sequencing data.
GRCh38: human reference genome build used for read alignment.
Smart-Seq3Xpress: plate-based single-cell RNA-seq library preparation method used to generate the dataset.
DNBSEQ: sequencing platform used to generate the raw sequencing reads.
FACS: fluorescence-activated cell sorting.
DAPI: viability dye used during cell sorting.
PBS: phosphate-buffered saline.
Data sources

Data sources:
This deposition contains raw sequencing data generated for this study, together with the associated zUMIs processing configuration and donor/sample metadata. It does not include raw data from external datasets.

Processed .h5ad count matrices and annotations, including integration with related external datasets, are provided separately.

Source code and software information:
Raw read processing was performed using zUMIs with the human GRCh38 reference genome. The deposited .yaml file contains the processing parameters used by zUMIs, including input FASTQ structure, barcode/UMI settings, reference genome settings, and output configuration.

Data quality and limitations:
The deposited FASTQ files contain raw sequencing reads and have not been filtered for downstream biological interpretation. Users should perform appropriate sequencing quality control, read-processing checks, and sample-level validation before reuse.