Published March 6, 2024 | Version v4
Dataset Open

AbSplice-DNA (hg38)

  • 1. Technical University of Munich
  • 2. University of California Irvine

Description

AbSplice-DNA predicts the probability that a variant causes aberrant splicing in a given tissue. See the publication: https://www.nature.com/articles/s41588-023-01373-3.
Here, we provide precomputed AbSplice-DNA scores for 49 human tissues and all possible SNVs genome-wide for hg38. This version contains 19,713 protein coding genes.

The folder 'AbSplice_DNA_hg38_snvs' contains all scores.
The folder 'AbSplice_DNA_hg38_snvs_high_scores' contains scores above 3 different cutoffs, which have approximately the same recalls as the high, medium and low cutoffs of SpliceAI: 

  • high cutoff (0.2),
  • medium cutoff (0.05),
  • low cutoff (0.01).

AbSplice scores are tissue-specific. In case users require a single score we recommend to use the maximum AbSplice score across tissues.

AbSplice-DNA scores can be computed from custom VCF files (including indels) with the python package 'absplice': https://github.com/gagneurlab/absplice

The uploaded files contain the following columns (for longer description see README of github repository of AbSplice):

  • Genomic coordinates of the variant:
    • chrom: Chromosome
    • pos: genomic position
    • ref: reference allele
    • alt: alternative allele
    • gene_id: Ensembl GeneID
  • AbSplice_DNA_{tissue}: AbSplice score for the given tissue
  • delta_logit_psi_{tissue}: MMSplice + SpliceMap score for a given tissue
  • delta_psi_{tissue}: MMSplice + SpliceMap + Ψ_ref score for a given tissue
  • splice_site_is_expressed_{tissue}: binary feature indicating if a splice site in the vicinity of the variant is expressed for a given tissue
  • delta_score: SpliceAI Delta score (maximum of Delta score (acceptor/donor gain/loss))
  • AbSplice_DNA_max: maximum AbSplice score across tissues for the given variant (this score is only provided in the files of the folder 'AbSplice_DNA_hg38_snvs_high_scores')

This dataset includes SpliceAI scores. The scores are free for academic and not-for-profit use; other use requires a commercial license from Illumina, Inc., see the GitHub repository of SpliceAI: https://github.com/Illumina/SpliceAI/tree/master

Files

AbSplice_DNA_hg38_snvs.zip

Files (24.2 GB)

Name Size Download all
md5:3b0fcf4b6c91c15f85ae489169dc4306
21.7 GB Preview Download
md5:f64a124bf7f061e888c4e32b2da64d33
2.5 GB Preview Download