Skip to content

thehyve/genetics-pipe

 
 

Repository files navigation

Genetics-pipe

Build Status

Everything is grch38 based. ETL pipeline used to integrate and generate the following tables:

  • gene to variant
  • disease to variant, and
  • disease to variant to gene

Howto build the pipeline

In order to use this pipeline the input data must follow an exact pattern described in these repositories

  • V2G Variant to gene data including tissues and cell-lines
  • V2D Variant to disease data

Build the code

The sbt assembly command will generate a fat-jar standalone jar that you can run locally or submit to a spark cluster. This jar already contains a default configuration file (src/main/resources/reference.conf) which includes present values. The configuration library follows Typesafe's documented configuration priority settings. In short, to add your own configuration file from the command line use -Dconfig.file=path/to/config-file. Any fields not present in that file will be resolved using the reference.conf file.

An example configuration file is provided under configuration/example.conf which lists all of the inputs which need to be specified. Updating this file is sufficient to run a genetics release.

Build a spark cluster on Google Cloud

To run the jar using Google's dataproc consult the script in ./scripts/run_cluster.sh which will start a cluster and submit a separate job for each step. The steps need to be executed in the correct order, so don't run asynchronously or change the specified order of the steps.

Run using a Google Dataproc Workflow

The simplest way to run everything is using the workflow defined in ./scripts/dataproc-workflow.sc. From within Intellij you can run it as a worksheet. The workflow defines each of the steps as a Dataproc job and runs it on a cluster. The cluster is automatically started and stopped, and jobs are parallelized where possible.

Alternatively you can run it from Ammonite but you need to ensure that all of the required jars are on the classpath. This can be done by executing the following command sbt Compile/fullClasspath/exportToAmmoniteScript && amm --predef target/scala-2.12/fullClasspath-Compile.sc. From here you can copy/paste the workflow into Ammonite.

Configuration

The configuration is defined in /src/main/resources/reference.conf. You can provide an external configuration with a subset of keys overwritten, for instance, only overwriting the keys specifying inputs.

The following inputs are required:

  • variant-index.raw
  • ensembl.lut
  • vep.homo-sapiens-cons-scores
  • interval.path
  • qtl.path
  • variant-gene.weights Deprecated as of 30 May 2022. Weights should be specified in the configuration field variant-gene.weights
  • variant-disease.studies
  • variant-disease.toploci
  • variant-disease.finemapping
  • variant-disease.ld
  • variant-disease.overlapping
  • variant-disease.coloc
  • variant-disease.trait_efo

For running internally within Open Targets consult the additional documentation.

Running steps

Step name Dependencies Output
variant-gene v2g
variant-index variant-index
dictionaries variant-index lut
search variant-index search
variant-disease variant-index v2d
variant-disease-coloc variant-index v2d_coloc
distance-nearest variant-index distance/canonical_tss
disease-variant-gene variant-disease d2v2g
scored-datasets variant-gene, disease-variant-gene v2g_by_overall, v2g_scored, d2v2g_by_overall, d2v2g_scored
manhattan l2g, scored-datasets, variant-disease-coloc manhattan

For a graphical representation of the dependencies see ./documentation/step_dependencies.puml

Variant to Gene table

Inputs

conf field notes
conf.vep.homoSapiensConsScores Notes taken directly from comment in code (might not be correct) load consequence table from file extracted from ensembl website https://www.ensembl.org/info/genome/variation/predicted_data.html#consequences and merged with OT eco scores table. We filter by only v2g_scores and get last token from the accession terms
conf.ensembl.lut Creates a gene index
conf.variantGene.path This is the output of the variant-gene step, so we assume there is a dependency between them.

Variant to disease table

TBD

Disease to variant to gene table

This table is a inner join between v2g and v2d

Search

This step produces simplified outputs of variant, gene and study in JSON-LD format suitable for ingestion into Elasticsearch. Only fields which are used for searching by the API are included.

Copyright

Copyright 2014-2018 Biogen, Celgene Corporation, EMBL - European Bioinformatics Institute, GlaxoSmithKline, Takeda Pharmaceutical Company and Wellcome Sanger Institute

This software was developed as part of the Open Targets project. For more information please see: http://www.opentargets.org

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Packages

 
 
 

Languages

  • Scala 94.8%
  • Shell 4.6%
  • Dockerfile 0.6%