lucapinello/crisprme-data

CRISPRme reference data & precomputed indexes

A fast, reliable mirror of the reference resources CRISPRme needs, plus precomputed indexes for the default references — so users can skip the slow raw-data download, the multi-hour variant enrichment, and the index build entirely.

Status: 🚧 work in progress — this repository is being populated. Layout and filenames below are the target structure; not all assets are uploaded yet.

Why this exists

A genuine from-scratch, genome-wide CRISPRme run was measured end-to-end and the cost is almost entirely data movement + enrichment, not the search:

• Downloading the 1000 Genomes VCFs from EBI ran at 1.4 MB/s (3 h); the UCSC genome download is similarly throttled. Hugging Face measured ~6.5× faster from the same machine.
• The variant enrichment stage was ~93 % of the ~13.5 h wall-clock (single-threaded per chromosome).

Hosting the data here (fast CDN) and — more importantly — shipping precomputed indexes removes the download and the enrichment/index-build for the common case, taking a genome-wide setup from many hours to minutes.

Target layout

crisprme-data/
├── genomes/
│   └── hg38/                      # per-chromosome FASTA (chr1.fa … chrX.fa)
├── vcfs/
│   ├── 1000G/                     # 1000 Genomes Phase 3 (GRCh38)
│   ├── HGDP/                      # Human Genome Diversity Project
│   └── unions/                    # combined cohorts (see "Default references")
├── annotations/                  # dhs+encode+gencode.hg38.bed, gencode.protein_coding.bed
├── pams/                          # e.g. 20bp-NGG-spCas9.txt, NNN motifs
├── samplesIDs/                    # per-dataset sample id lists
└── indexes/                       # PRECOMPUTED genome_library + enriched indexes
    ├── NGG_1000G+HGDP+TOPMed+AllofUs/
    ├── NNN_1000G+HGDP/            # backward-compatible with the original CRISPRme
    └── NGG_pangenome2.0/

Default references (shipped as precomputed indexes)

Two headline references (highlighted in the manuscript):

1. Unified large-cohort panel — 1000G + HGDP + TOPMed + All of Us.
2. Pangenome 2.0 VCF.

Plus an NNN + 1000G+HGDP configuration for backward compatibility with the original CRISPRme, and default PAMs NNN and NGG.

Usage

CRISPRme's setup will be able to pull these resources from this repository instead of the original (slow) upstream hosts. Direct download also works, e.g.:

# whole repo (large) — prefer fetching only what you need
hf download pinellolab/crisprme-data --repo-type dataset --local-dir crisprme-data

License & citation

Reference data are redistributed under the terms of their original sources (1000 Genomes, HGDP, GENCODE/ENCODE, UCSC hg38, etc.). CRISPRme is AGPL-3.0.

If you use CRISPRme, please cite: Cancellieri S, et al. Human genetic diversity alters off-target outcomes of therapeutic gene editing. Nat Genet 55, 34–43 (2023). doi:10.1038/s41588-022-01257-y

Project: https://github.com/pinellolab/CRISPRme