VIBE

A tool to generate prioritized genes using phenotype information. See publication K. Joeri van der Velde, Sander van den Hoek et al (2020), Advanced Genetics 1(1).

Web app

• Available at: https://vibe.molgeniscloud.org/
• Source code: https://github.com/molgenis/molgenis-app-vibe

Benchmark

• Available at: https://github.com/molgenis/vibe-suppl

Quickstart

• Download vibe .jar file
• Download and extract HDT (sha256 checksum)
• Download the HPO.owl
• Make sure you have Java 8 or higher
• Open a terminal and run VIBE. java -jar vibe-with-dependencies-<version>.jar -d -t vibe-<db-version>.hdt -o results.tsv -p HP:0002996 -p HP:0001377

Detailed instructions

Requirements

• Java 8 or higher
• A local HDT dataset
• Human Phenotype Ontology (HPO)

Usage

java -jar vibe-with-dependencies.jar [-h] [-v] [-d] [-f] -t <FILE> -w <FILE> [-n <NAME> -m <NUMBER>] [-o <FILE>] [-l] [-u] -p <HPO ID> [-p <HPO ID>]...

IMPORTANT: Do keep this in mind. Especially when using -n.

Examples

Using only the user-defined phenotypes:

java -jar vibe-with-dependencies-<version>.jar -d -t vibe-<db-version>.hdt -w hp.owl -o results.tsv -p HP:0002996 -p HP:0001377

Using the user-defined phenotypes and phenotypes that are related to them with a maximum distance of 1:

java -jar vibe-with-dependencies-<version>.jar -d -t vibe-<db-version>.hdt -w hp.owl -n distance -m 1 -o results.tsv -p HP:0002996

Using the user-defined phenotypes and their (grand)children with a maximum distance of 2:

java -jar vibe-with-dependencies-<version>.jar -d -t vibe-<db-version>.hdt -w hp.owl -n children -m 2 -o results.tsv -p HP:0002996

Output format

There are currently 2 options for the output. By default the output will look something like:

gene (NCBI) gene symbol (HGNC)  highest GDA score   diseases (UMLS) with sources per disease
1311    COMP    1.0 C0410538 (1.0):10405447,10753957,10852928,11084047,11691584,11745002,11746044,11746045,11782471,11891674,11968079,12479386,12483304,12483437,12768438,12792737,12819015,14580238,15094116,15183431,15266613,15337766,15551305,15579310,15694129,15756302,15880723,16199550,16514635,16520029,17200202,17307347,17394206,17570134,17579668,17588960,18193163,18546327,19762713,20301660,20578249,20819661,20936634,21042783,21599986,21922596,22006726,23562786,24194321,24595329,24892720,24997222,29104872,7670471,7670472,9021009,9184241,9188668,9388247,9452026,9452063,9463320,9632164,9749943,9880218,9887340,9921895,9923655|C1838280 (0.74):10405447,11084047,11565064,11782471,12483304,12819015,14684695,20301302,21922596,7670472,9021009,9184241,9452026,9463320,9887340,9921895,9923655|C0026760 (0.6):10364514,10655510,11084047,11479597,11528506,11565064,11782471,11968079,12479386,12483304,12819015,15183431,15337766,15694129,15756302,15880723,16199550,16514635,17133256,17200202,17570134,18193163,18682400,19808781,20578249,20936634,21922596,24595329,24997222,7670472,9184241,9463320,9921895,9923655|C0013336 (0.13):11691584,12483304,15472220|C1867103 (0.1)|C0029422 (0.03):11891674,12768438,17579668
4010    LMX1B   1.0 C0027341 (1.0):10425280,10537763,10571942,10660670,10767331,10854116,11668639,11956244,11978876,12215822,12792813,12819019,15562281,15638822,15774843,15785774,15928687,16825280,17166916,17431898,17515884,17657578,17710881,18414507,18535845,18538102,18562181,18595794,18634531,18952915,19147669,19222527,19721866,20199424,20531206,20568247,21184584,21850167,22211385,23687361,24042019,24720768,25380522,25898926,26380986,26560070,28335748,9590287,9590288,9618165,9664684,9837817|C0029422 (0.11):10767331|C1867103 (0.1)|C1867439 (0.1)

Here, each row represents a gene. Each row consists out of several (tab-seperated) fields. First is the NCBI gene id, then the highest found gene-disease association score found in DisGeNET for that gene, and finally all found associated diseases with the gene-disease association score specific for that gene-disease combination followed by the evidence IDs (pubmed if no URL is shown) for it.

Alternatively, the option -l can be added when running the tool. If this is done, the output will only contain the genes (separated by comma's). The example output above would look like:

1311,4010

The -u option changes the output to use URI's instead of ID's for certain fields. As this strongly increases the output size, in general it is not advised to use this option.

For developers

Below are the instructions on how to build the app. For instructions on how to create the HDT, please check here.

Requirements

• Apache Maven (download and install)

Preperations

Before building/testing, be sure the needed test resources are downloaded. This can be done by running TestsPreprocessor.sh. When checking out a commit that uses a different version of the test resources (as defined by the pom.xml properties vibe-database.version and hpo-owl.tag), be sure to re-run this script so that the correct resources are used for testing.

Building executables

1. Run mvn clean install to create all needed files.
   • An überjar for the command-line tool can be found at vibe-cli/target/vibe-with-dependencies-<version>.jar .
   • Checksums are available in the <sub-module>/target/checksums/ folders for the following formats:
     • MD5
     • SHA-256
     • SHA-512
     • SHA3-512

Branches

• release -> release branches (f.e. release/5.0)
• feature -> a branch for a new feature (f.e. feature/hdt)
• fix -> implementing a fix (f.e. fix/invalid_hgnc)
• chore -> simple maintenance tasks (f.e. chore/readme_update)
• backmerge -> merging back fixes on release branches back to master (f.e. backmerge/5.0)

Notes

• The different modules partly use the same resources. To not create duplicate data, these are stored in the shared_resources/shared folder. When adding new test resources to a module, ensure that there is no src/test/resources/shared directory.
• In case tests fail on Jenkins due to Jenkins-specific reasons (f.e. due to requiring to change a file permission of a test resource), the JUnit tag skipOnJenkins can be used so that these test still occur when building locally.