PolyGenius

Resources

The external resources PolyGenius resolves on your behalf. You rarely need this page: every asset here is fetched, verified and cached automatically the first time an analysis asks for it. Direct links are for air-gapped machines, shared caches, and mirrors.

Reference panels

LD reference genotypes used by LDpred2, lassosum2 and PRS-CS, and by clumping in C+T. 1000 Genomes super-populations, MAF>0.1, SNP-only, multi-allelics binarized. Published as hg19 PLINK 2 pfiles; other builds are lifted on demand.

workspace$catalogs$referencePanels
manifest.csv
# Inspect what is available, then resolve one — downloaded and cached on first use
workspace$catalogs$referencePanels$view()
panel <- workspace$catalogs$referencePanels$get("EUR", build = "hg19")

# Other builds are derived on demand by the catalog's liftover rule
panel38 <- workspace$catalogs$referencePanels$get("EUR", build = "hg38")
  • AFR
    hg19 · pfile · AFR fileset

    1000G AFR super-population.

    Download
  • AMR
    hg19 · pfile · AMR fileset

    1000G AMR super-population.

    Download
  • EAS
    hg19 · pfile · EAS fileset

    1000G EAS super-population.

    Download
  • EUR
    hg19 · pfile · EUR fileset

    1000G EUR super-population.

    Download
  • SAS
    hg19 · pfile · SAS fileset

    1000G SAS super-population.

    Download

Variant spaces

Curated variant subsets used to restrict a panel or a PGS library to a common, well-imputed backbone.

workspace$catalogs$variantSpaces
manifest.csv
workspace$catalogs$variantSpaces$view()
space <- workspace$catalogs$variantSpaces$get("common20k", build = "hg38")
  • common20k
    hg19 · 20,000 variants

    Default common 20k variant space.

    Download
  • common20k
    hg38 · 20,000 variants

    Default common 20k variant space.

    Download
  • HapMap3+
    hg19 · 1,444,192 variants

    LDpred2 HapMap3+ variant space.

    Download
  • HapMap3+
    hg38 · 1,443,186 variants

    LDpred2 HapMap3+ variant space.

    Download

Liftover chains

UCSC chain files for converting positions between genome builds. Resolved automatically whenever a model and a genotype set disagree on build.

workspace$catalogs$liftoverChains
manifest.csv
workspace$catalogs$liftoverChains$view()
chain <- workspace$catalogs$liftoverChains$get(from = "hg19", to = "hg38")

LD blocks

Approximately-independent LD block boundaries, used to partition the genome for block-wise estimators. hg38 sets are lifted from the published hg19 boundaries.

workspace$catalogs$LDblocks
manifest.csv
workspace$catalogs$LDblocks$view()
blocks <- workspace$catalogs$LDblocks$get(population = "EUR", build = "hg38")
  • AFR · hg19
    Berisa & Pickrell 2016 (ldetect-data, fourier_ls-all.bed)
    Download
  • AFR · hg38
    Berisa & Pickrell 2016 (ldetect-data), liftOver hg19->hg38
    Download
  • EAS · hg19
    Berisa & Pickrell 2016 (ldetect-data, fourier_ls-all.bed)
    Download
  • EAS · hg38
    Berisa & Pickrell 2016 (ldetect-data), liftOver hg19->hg38
    Download
  • EUR · hg19
    Berisa & Pickrell 2016 (ldetect-data, fourier_ls-all.bed)
    Download
  • EUR · hg38
    Berisa & Pickrell 2016 (ldetect-data), liftOver hg19->hg38
    Download

GCTB

Required only by SBayes-family algorithms. Prebuilt binaries are Linux x86_64 only; elsewhere, compile from source and register the path.

workspace$setup
manifest.csv
workspace$setup$get("gctb")
workspace$setup$install(gctb = "/path/to/gctb")