PIASO-data

Genome reference files and tutorial datasets for PIASO (Precise Integrative Analysis of Single-cell Omics).

Tutorial Datasets

All datasets are hosted on Zenodo under concept DOI 10.5281/zenodo.19699638 (always resolves to the latest version; current version: record 22012620).

The smallest dataset is E18 Nuclei 5K v3.1 (e18_v3_nuclei, 19.3 MB): use it for tests, examples and first runs, and reach for the larger ones only when the tutorial calls for them.

Dataset	Species	Cells	Format	Size	Tutorial	Reference
SEA-AD MTG 20K (raw)	Human	20,000	h5ad	1.8 GB	PIASO intro, plotting, color palettes	Gabitto et al. Nat Neurosci (2024)
Adult Cortex Multiome RNA	Mouse	17,412	h5ad	2.5 GB	GDR, marker gene prediction, markerDB	Bravo González-Blas et al. Nat Methods (2023)
10K Mouse Brain GEM-X v4	Mouse	11,357	10x h5	65.5 MB	RNA pipeline (single sample)	10x Genomics
E18 Neurons 10K v3	Mouse	~10,000	10x h5	45.4 MB	RNA pipeline (multi-sample)	10x Genomics
E18 Nuclei 5K v3.1	Mouse	~5,000	10x h5	19.3 MB	RNA pipeline (multi-sample)	10x Genomics
E18 Neurons 10K GEM-X v4	Mouse	~10,000	10x h5	64.5 MB	RNA pipeline (multi-sample)	10x Genomics
PBMC snMultiome SAN1	Human	3,545	10x h5	73.2 MB	PBMC pipeline (multi-sample)	De Rop et al. Nat Biotechnol (2024)
PBMC snMultiome SAN2	Human	4,360	10x h5	83.7 MB	PBMC pipeline (single + multi)	De Rop et al. Nat Biotechnol (2024)
PIASOmarkerDB Allen Immune	Human	—	CSV	115 KB	PIASOmarkerDB API	Gong et al. Nature (2025)

Cytome datasets

Ready-to-stream .cytome files — open them directly, no conversion step, constant memory at any scale:

Dataset (registry name)	Species	Cells	Size	Tutorial	Reference
sea_ad_mtg_20k_cytome	Human	20,000	269 MB	cytome basics	Gabitto et al. Nat Neurosci (2024)
adult_cortex_multiome_rna_cytome	Mouse	17,412	192 MB	GDR on cytome	Bravo González-Blas et al. Nat Methods (2023)
allen_devvis_rna	Mouse	200,061	1.4 GB	GDR at scale	Gao et al. Nature (2025)
humandevcx_38_rna	Human	213,090	1.1 GB	large-scale workflows	Wang et al. Nature (2025)
humanlifespan_pfc_rna	Human	1,501,089	25.7 GB	million-cell streaming	Catching et al. Cell Reports (2026)

Each file stores raw UMI counts (RNA_counts, verified integer at conversion) plus cell annotations from the source atlas.

import piaso

# Stream a 200k-cell dataset without loading it into memory
ds = piaso.data.load_dataset("allen_devvis_rna", return_type="cytome")

Programmatic access

import piaso

# List all available datasets
piaso.data.list_datasets()

# Download and load a dataset
adata = piaso.data.load_dataset("sea_ad_mtg_20k")

# Just download (returns local path)
path = piaso.data.fetch_dataset("mouse_brain_10k_gemx")

Files are cached in ~/.piaso/data/datasets/ by default. To use a different location (PIASO ≥ 1.2.1), any of these works — most specific wins:

piaso.data.fetch_dataset("allen_devvis_rna", data_dir="/big/disk/piaso")  # per call
piaso.settings.data_dir = "/big/disk/piaso"                               # per session
# or per machine:  export PIASO_DATA_DIR=/big/disk/piaso

The dataset registry (datasets.json) is fetched from this repo and cached locally for 24 h; piaso.data.refresh_registry() forces an update.

Genome References

hg38 (Human)

File	Description
hg38_genes.bed	Gene body BED (chrom, start, end, gene_name, score, strand)
hg38_promoterSet.bed	Promoter regions BED
GRCh38-cCREs.CTCF-only.bed	ENCODE cCREs CTCF-only sites
hg38.chrom.sizes	Chromosome sizes
hg38_transcript_tss.bed	Transcript TSS positions

mm10 (Mouse)

File	Description
mm10_genes.bed	Gene body BED (chrom, start, end, gene_name, score, strand)
mm10_promoterSet.bed	Promoter regions BED
mm10-cCREs.CTCF-only.bed	ENCODE cCREs CTCF-only sites
mm10.chrom.sizes	Chromosome sizes
mm10_transcript_tss.bed	Transcript TSS positions

Mmul_10 (Rhesus macaque) and mCalJac1 (Common marmoset)

File	Description
<genome>_genes.bed	Gene body BED (chrom, start, end, gene_name, score, strand)
<genome>_promoterSet.bed	Promoter regions BED (TSS −400 / +200 bp, strand-aware)
<genome>.chrom.sizes	Chromosome sizes
<genome>_transcript_tss.bed	Transcript TSS positions

No cCRE file: ENCODE SCREEN publishes candidate cis-regulatory elements for human and mouse only.

Provenance. Both were derived from the cellranger-arc references we map our own macaque and marmoset data against, not from NCBI directly, so the chromosome names in these files are the names your fragments will carry if you use the same references. The references were built as follows:

	Mmul_10	mCalJac1
NCBI RefSeq assembly	GCF_003339765.1 (rheMac10, Mmul_10)	GCF_011100555.1 (mCalJa1.2.pat.X)
Annotation	NCBI Macaca mulatta Annotation Release 103	NCBI RefSeq GCF_011100555.1-RS_2023_03
Sequence names	RefSeq accessions renamed to chr1..chr20, chrX, chrY, chrM; unplaced scaffolds keep their chrUn_*_random names	RefSeq accessions renamed to chr1..chr22, chrX, chrY; unlocalized scaffolds keep their scaffold names
Mitochondrion	chrM (NC_005943.1) added by hand	none — the assembly has no mitochondrial contig
GTF filter	cellranger-arc mkgtf keeping protein_coding, lncRNA, antisense, IG_* and TR_* biotypes	same
Motifs	JASPAR 2022 CORE vertebrates non-redundant, headers as NAME_ID	same file
Built with	cellranger-arc 2.0.2 mkref	cellranger-arc 2.0.2 mkref

The prebuilt references (about 30 GB each, with the BWA and STAR indices) are mirrored on Google Drive for anyone who wants to map with exactly the same reference: Mmul_10.tar.gz and mCalJac1.tar.gz. They are usable directly with cellranger-arc count (multiome) and, for the GEX and ATAC halves, with cellranger count and cellranger-atac count. Treat the Drive copies as a convenience mirror: the BED files in this repository are the versioned, diffable artefacts PIASO actually reads.

Because the marmoset assembly has no mitochondrial contig, any mitochondrial-fraction QC on mCalJac1 data returns zero for every cell — that describes the assembly, not the cells. PIASO records this in piaso.data.GENOME_FEATURES["mCalJac1"]["has_chrM"].

Building the files for another species

The four BED files come from any cellranger or cellranger-arc reference directory, so you can generate them for your own genome with the same conventions as the shipped ones:

import piaso
piaso.data.makeGenomeFiles("/path/to/cellranger_ref", genome="MyGenome",
                           outdir="~/.piaso/data/MyGenome")

Build from the reference you mapped against rather than from the source annotation: the reference carries whatever chromosome renaming was done when it was made, and BED files whose names disagree with the fragments do not raise an error, they just find nothing.

Genome usage

import piaso

# Auto-download genome files (fetches from this repo)
piaso.data.fetch_genome("hg38")
piaso.data.fetch_genome("mm10")
piaso.data.fetch_genome("Mmul_10")
piaso.data.fetch_genome("mCalJac1")

Manual genome installation

mkdir -p ~/.piaso/data
tar -xzf hg38.tar.gz -C ~/.piaso/data/
tar -xzf mm10.tar.gz -C ~/.piaso/data/

References

• Gabitto, M.I., Travaglini, K.J., Rachleff, V.M. et al. Integrated multimodal cell atlas of Alzheimer's disease. Nat Neurosci 27, 2366–2383 (2024). DOI: 10.1038/s41593-024-01774-5
• Bravo González-Blas, C., De Winter, S., Hulselmans, G. et al. SCENIC+: single-cell multiomic inference of enhancers and gene regulatory networks. Nat Methods 20, 1355–1367 (2023). DOI: 10.1038/s41592-023-01938-4
• De Rop, F.V., Hulselmans, G., Flerin, C. et al. Systematic benchmarking of single-cell ATAC-sequencing protocols. Nat Biotechnol 42, 916–926 (2024). DOI: 10.1038/s41587-023-01881-x
• Gao, Y., van Velthoven, C.T.J., Lee, C. et al. Continuous cell-type diversification in mouse visual cortex development. Nature 647, 127–142 (2025). DOI: 10.1038/s41586-025-09644-1
• Wang, L., Wang, C., Moriano, J.A. et al. Molecular and cellular dynamics of the developing human neocortex. Nature 647, 169–178 (2025). DOI: 10.1038/s41586-024-08351-7
• Catching, A., Weller, C.A., Hu, F. et al. Single-nucleus multiome analysis in the human prefrontal cortex identifies gene expression and cis-regulatory elements associated with aging. Cell Reports 45, 117110 (2026). DOI: 10.1016/j.celrep.2026.117110
• Gong, Q., Sharma, M., Glass, M.C. et al. Multi-omic profiling reveals age-related immune dynamics in healthy adults. Nature 648, 696–706 (2025). DOI: 10.1038/s41586-025-09686-5

License

Genome reference files are derived from public annotations (UCSC, ENCODE, NCBI RefSeq). Tutorial datasets are redistributed under CC BY 4.0 with attribution to original sources.