Skip to main content

SpliceAI: A deep learning-based tool to identify splice variants

Project description

SpliceAI: A deep learning-based tool to identify splice variants

release license downloads

This package annotates genetic variants with their predicted effect on splicing, as described in Jaganathan et al, Cell 2019 in press.

Update: The annotations for all possible substitutions, 1 base insertions, and 1-4 base deletions within genes are available here for download.

Installation

The simplest way to install SpliceAI is through pip or conda:

pip install spliceai
# or
conda install -c bioconda spliceai

Alternately, SpliceAI can be installed from the github repository:

git clone https://github.com/Illumina/SpliceAI.git
cd SpliceAI
python setup.py install

SpliceAI requires tensorflow>=1.2.0, which is best installed separately via pip or conda (see the TensorFlow website for other installation options):

pip install tensorflow
# or
conda install tensorflow

Usage

SpliceAI can be run from the command line:

spliceai -I input.vcf -O output.vcf -R genome.fa -A grch37
# or you can pipe the input and output VCFs
cat input.vcf | spliceai -R genome.fa -A grch37 > output.vcf

Required parameters:

  • -I: Input VCF with variants of interest.
  • -O: Output VCF with SpliceAI predictions ALLELE|SYMBOL|DS_AG|DS_AL|DS_DG|DS_DL|DP_AG|DP_AL|DP_DG|DP_DL included in the INFO column (see table below for details). Only SNVs and simple INDELs (REF or ALT is a single base) within genes are annotated. Variants in multiple genes have separate predictions for each gene.
  • -R: Reference genome fasta file. Can be downloaded from GRCh37/hg19 or GRCh38/hg38.
  • -A: Gene annotation file. Can instead provide grch37 or grch38 to use GENCODE V24 canonical annotation files included with the package. To create custom gene annotation files, use spliceai/annotations/grch37.txt in repository as template.

Optional parameters:

  • -D: Maximum distance between the variant and gained/lost splice site (default: 50).
  • -M: Mask scores representing annotated acceptor/donor gain and unannotated acceptor/donor loss (default: 0).

Details of SpliceAI INFO field:

ID Description
ALLELE Alternate allele
SYMBOL Gene symbol
DS_AG Delta score (acceptor gain)
DS_AL Delta score (acceptor loss)
DS_DG Delta score (donor gain)
DS_DL Delta score (donor loss)
DP_AG Delta position (acceptor gain)
DP_AL Delta position (acceptor loss)
DP_DG Delta position (donor gain)
DP_DL Delta position (donor loss)

Delta score of a variant, defined as the maximum of (DS_AG, DS_AL, DS_DG, DS_DL), ranges from 0 to 1 and can be interpreted as the probability of the variant being splice-altering. In the paper, a detailed characterization is provided for 0.2 (high recall), 0.5 (recommended), and 0.8 (high precision) cutoffs. Delta position conveys information about the location where splicing changes relative to the variant position (positive values are downstream of the variant, negative values are upstream).

Examples

A sample input file and the corresponding output file can be found at examples/input.vcf and examples/output.vcf respectively. The output T|RYR1|0.00|0.00|0.91|0.08|-28|-46|-2|-31 for the variant 19:38958362 C>T can be interpreted as follows:

  • The probability that the position 19:38958360 (=38958362-2) is used as a splice donor increases by 0.91.
  • The probability that the position 19:38958331 (=38958362-31) is used as a splice donor decreases by 0.08.

Similarly, the output CA|TTN|0.07|1.00|0.00|0.00|-7|-1|35|-29 for the variant 2:179415988 C>CA has the following interpretation:

  • The probability that the position 2:179415981 (=179415988-7) is used as a splice acceptor increases by 0.07.
  • The probability that the position 2:179415987 (=179415988-1) is used as a splice acceptor decreases by 1.00.

Contact

Kishore Jaganathan: kishorejaganathan@gmail.com

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

spliceai-1.3.tar.gz (16.7 MB view hashes)

Uploaded Source

Built Distribution

spliceai-1.3-py2.py3-none-any.whl (16.7 MB view hashes)

Uploaded Python 2 Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page