KipoiSplice/4

Authors: Ziga Avsec , Roman Kreuzhuber

License: MIT

Contributed by: Ziga Avsec , Roman Kreuzhuber

Postprocessing: None

Trained on: ClinVar (release 2018-04-29) variants in the range [-40nt, 10nt] around the splicing acceptor or variants in the range [-10, 10] nt around the splice donor of a protein coding gene. Only variants labelled 'Pathogenic' or 'Benign' were used. Data from all chromosomes was used for training.

Source files

Kipoi ensemble model without conservation. The used models are - MaxEntScan/3prime, MaxEntScan/5primet, labranchor, HAL

Create a new conda environment with all dependencies installed

kipoi env create KipoiSplice/4
source activate kipoi-KipoiSplice__4

Test the model

kipoi test KipoiSplice/4 --source=kipoi

Make a prediction

kipoi get-example KipoiSplice/4 -o example
kipoi predict KipoiSplice/4 \
  --dataloader_args='{"fasta_file": "example/fasta_file", "gtf_file": "example/gtf_file", "num_workers": 1, "vcf_file": "example/vcf_file"}' \
  -o '/tmp/KipoiSplice|4.example_pred.tsv'
# check the results
head '/tmp/KipoiSplice|4.example_pred.tsv'

Create a new conda environment with all dependencies installed

kipoi env create KipoiSplice/4
source activate kipoi-KipoiSplice__4

Get the model

import kipoi
model = kipoi.get_model('KipoiSplice/4')

Make a prediction for example files

pred = model.pipeline.predict_example(batch_size=4)

Use dataloader and model separately

# Download example dataloader kwargs
dl_kwargs = model.default_dataloader.download_example('example')
# Get the dataloader and instantiate it
dl = model.default_dataloader(**dl_kwargs)
# get a batch iterator
batch_iterator = dl.batch_iter(batch_size=4)
for batch in batch_iterator:
    # predict for a batch
    batch_pred = model.predict_on_batch(batch['inputs'])

Make predictions for custom files directly

pred = model.pipeline.predict(dl_kwargs, batch_size=4)

Get the model

library(reticulate)
kipoi <- import('kipoi')
model <- kipoi$get_model('KipoiSplice/4')

Make a prediction for example files

predictions <- model$pipeline$predict_example()

Use dataloader and model separately

# Download example dataloader kwargs
dl_kwargs <- model$default_dataloader$download_example('example')
# Get the dataloader
dl <- model$default_dataloader(dl_kwargs)
# get a batch iterator
it <- dl$batch_iter(batch_size=4)
# predict for a batch
batch <- iter_next(it)
model$predict_on_batch(batch$inputs)

Make predictions for custom files directly

pred <- model$pipeline$predict(dl_kwargs, batch_size=4)

Get the docker image

docker pull kipoi/kipoi-docker:kipoisplice-slim

Get the full sized docker image

docker pull kipoi/kipoi-docker:kipoisplice

Get the activated conda environment inside the container

docker run -it kipoi/kipoi-docker:kipoisplice-slim

Test the model

docker run kipoi/kipoi-docker:kipoisplice-slim kipoi test KipoiSplice/4 --source=kipoi

Make prediction for custom files directly

# Create an example directory containing the data
mkdir -p $PWD/kipoi-example 
# You can replace $PWD/kipoi-example with a different absolute path containing the data 
docker run -v $PWD/kipoi-example:/app/ kipoi/kipoi-docker:kipoisplice-slim \
kipoi get-example KipoiSplice/4 -o /app/example 
docker run -v $PWD/kipoi-example:/app/ kipoi/kipoi-docker:kipoisplice-slim \
kipoi predict KipoiSplice/4 \
--dataloader_args='{'fasta_file': '/app/example/fasta_file', 'gtf_file': '/app/example/gtf_file', 'num_workers': 1, 'vcf_file': '/app/example/vcf_file'}' \
-o '/app/KipoiSplice_4.example_pred.tsv' 
# check the results
head $PWD/kipoi-example/KipoiSplice_4.example_pred.tsv

Install apptainer

https://apptainer.org/docs/user/main/quick_start.html#quick-installation-steps

Make prediction for custom files directly

kipoi get-example KipoiSplice/4 -o example
kipoi predict KipoiSplice/4 \
--dataloader_args='{"fasta_file": "example/fasta_file", "gtf_file": "example/gtf_file", "num_workers": 1, "vcf_file": "example/vcf_file"}' \
-o 'KipoiSplice_4.example_pred.tsv' \
--singularity 
# check the results
head KipoiSplice_4.example_pred.tsv

Schema

Inputs

Single numpy array

Name: None

Shape: (12,)

Doc: Model predictions of MaxEntscan, HAL and labranchor

Targets

Single numpy array

Name: None

Shape: (2,)

Doc: Pathogenicity score. 0th index represents the probability of being benign and 1st index represents the probability for being pathogenic.

Dataloader

Defined as: .

Doc: This model first predicts the effect of variants using 4 splicing models (MaxEntScan/3prime, MaxEntScan/5prime, HAL and labranchor) and then integrates those predictions using logistic regression. The model has been trained to distinguish between pathogenic and benign variants in ClinVar.

Authors: Ziga Avsec , Roman Kreuzhuber

Type: PreloadedDataset

License: MIT

Arguments

batch_size (optional): batch size to use with all the models

fasta_file : reference genome fasta file

gtf_file : path to the GTF file required by the models (Ensemble)

num_workers (optional): number of workers to use for each model

tmpdir (optional): path to the temporary directory where to store the predictions

vcf_file : Path to the input vcf file

Model dependencies

conda:

pip=20.3.3

pip:

scikit-learn==0.22.2.post1
sklearn-pandas==1.8.0
tensorflow==1.13.1
numexpr==2.6.2

Dataloader dependencies

conda:

bioconda::pysam
bioconda::maxentpy
bioconda::pybedtools
bioconda::cyvcf2
pandas
h5py
numpy
python=3.5
attrs=17.4.0

pip:

pyvcf
intervaltree
joblib
scikit-learn
sklearn-pandas
kipoi==0.6.30
kipoi_utils==0.7.2
kipoi_veff
tqdm
tensorflow>=1.0.0
keras==2.2.4
protobuf==3.19.4