The Exomiser - Web Code

##Setup

This is a Spring Boot jar application meaning it runs using its own embedded Tomcat server. In this case we're using the embedded version of the H2 database as this performs well, in particular because the data is local to the machine.

Once built, to run the application will need to:

1. Have the exomiser-cli data directory somewhere visible to the machine you're running the server from.
2. Copy the ${project.name}-${project.version}.jar to the machine you want to run it from.
3. Make sure you have the application.properties file alongside the jar file. The defaults can be found in /src/main/resources/application.properties:

   
  #due to hardware and page timeout constraints
  #we need to limit the maximum number of variants which will be analysed
  maxVariants=100000
  #max genes especially hits the RAM usage for rendering a page so this is limited
  maxGenes=200
  
  exomiser.data-directory=/full/system/path/to/exomiser/data

##Running Launch the application by intoning the incantation thus:

  java -jar -XX:+UseG1GC -Xms8G -Xmx8G ${project.name}-${project.version}.jar

The port will default to 8080, but it can be changed by specifying it either on the command line:

  java -jar -XX:+UseG1GC -Xms8G -Xmx8G ${project.name}-${project.version}.jar --server.port=8090

or by adding it to the application.properties

  server.port=8090

Further details can be found on the Spring Boot Deployment reference.

Check all is well by visiting the submission page:

  http://localhost:8080/exomiser/submit

The submission page now contains a warning, shown by default, alerting users not to input sensitive clinical data. If the instance is running in a clinically-compliant setting, the banner can be removed by setting the field exomiser.web.clinical-instance to true in the application.properties file:

exomiser.web.clinical-instance=true

or override this by launching the application using the command-line option

--exomiser.web.clinical-instance=true

Running the Docker image

To run the image you will need the standard exomiser directory layout to mount as separate volumes and supply an application.properties file or environmental variables to point to the data required e.g.

docker run \
-v "/data/exomiser-data:/exomiser-data" \
-v "/opt/exomiser/exomiser-config/:/exomiser" \
-p 8080:8080 \
exomiser/exomiser-web:latest \
--spring.config.location=/exomiser/application.properties

or using Spring configuration arguments instead of the application.properties:

docker run -v "/data/exomiser-data:/exomiser-data" \
-p 8080:8080 \
exomiser/exomiser-web:latest \
--exomiser.data-directory=/exomiser-data \
--exomiser.hg19.data-version=2109 \
--exomiser.hg19.variant-white-list-path=2109_hg19_clinvar_whitelist.tsv.gz \
--exomiser.phenotype.data-version=2109

Here the contents of /opt/exomiser/exomiser-config is simply the application.properties file.

$ tree /opt/exomiser/exomiser-config/
exomiser-config/
└──application.properties

and the application.properties contents for hg38:

exomiser.data-directory=/exomiser-data

## hg38 config
exomiser.hg38.data-version=2109
#exomiser.hg38.remm-path=${exomiser.data-directory}/remm/ReMM.v0.3.1.post1.hg38.tsv.gz
#exomiser.hg38.local-frequency-path=${exomiser.data-directory}/local/local_frequency_test_hg38.tsv.gz
exomiser.hg38.variant-white-list-path=${exomiser.hg38.data-version}_hg38_clinvar_whitelist.tsv.gz

## phenotype config
exomiser.phenotype.data-version=2109

Resizing the JVM

Running a WGS analysis can take a few GB of RAM, depending on the size of the sample in question. If you need to increase the max memory of the JVM, include the following environment variable:

-e JAVA_TOOL_OPTIONS="-Xmx8G"