Target - Adverse Reaction Database Integrated Search (T-ARDIS)

Cristiano Galletti, Patricia Mirela Bota, Baldo Oliva, Narcis Fernandez-Fuentes, Mining drug–target and drug–adverse drug reaction databases to identify target–adverse drug reaction relationships, Database, Volume 2021, 2021, baab068, https://doi.org/10.1093/database/baab068

This repository contains all the file necessary to download, compute and statistically validate the pairwise interaction between Drug side effects and Drug Targets.

T-ARDIS is now avaiable as webservice at http://www.bioinsilico.org/T-ARDIS/

In order to make it works a couple of pre-processing steps are necessary.

1. Set up a postgresql database where to run the FAERS cleaning procedure. It's necessary just create the server, all other procedure, like the database creation are done automatically in the bash script

2. This procedure has been run using the Anaconda packages, So it's suggested to download the latest Anaconda repository or set up a python environment with the packages in the requirements file

3. Download an updated version of ATHENA vocabularies. These vocabularies are used to map and retrieve unique DRUG names in the FAERS procedure. The total number of used vocabularies are 23, included the SNOWMED and MEDDRA dictionaries.

• To download that you need to be registered in the ATHENA website http://athena.ohdsi.org and have a license for the access the MEDDRA dictionary, the latter can be requested registering in the UMLS website.
• Once the registering procedures are completed, to download the dictionaries just go again to the ATHENA website at http://athena.ohdsi.org/vocabulary/list, add to the already selected databases the MEDDRA and click on "download vocabularies".
• Save and extract the files in a folder called "athena" in the same parent folder in which are saved all the other repository files.

4. Download the latest MEDDRA release on https://www.meddra.org/software-packages (you need to login with the Meddra credential and request unzip code from https://apps.meddra.org/selfservice/get_unzip_password2.aspx to extract the files). This is needed for a supplementary cleaning procedure. Extract the zip file in the same working directory.

5. Launch the run_file.sh script

• The script will start downloading the updated files of the different DRUG - SIDE EFFECT databases:

  • FAERS (https://www.fda.gov/drugs/questions-and-answers-fdas-adverse-event-reporting-system-faers/fda-adverse-event-reporting-system-faers-latest-quarterly-data-files)

  • MEDEFFECT (https://www.canada.ca/en/health-canada/services/drugs-health-products/medeffect-canada/adverse-reaction-database/medeffect-canada-caveat-privacy-statement-interpretation-data-search-canada-vigilance-adverse-reaction-online-database.html)

  • OFFSIDE (http://tatonettilab.org/offsides/)

  • SIDER_4.1 (http://sideeffects.embl.de/)

• And the updated versions of the DRUG - TARGET database:

  • DRUG TARGET COMMONS (https://drugtargetcommons.fimm.fi/)
  • STITCH 5.0 (http://stitch.embl.de/)

6. The procedure will standardize, clean and relate the information in the databases to obtain a list of pairwise relationships between the Drug's targets and side effects exploiting the drug name as bridge. In order to confirm the relationship the results have been statistically validated using the Fisher exact test as explained in the paper of Kuhn et al. (http://europepmc.org/article/MED/23632385)

At the end of the procedure, following the example of [https://pubmed.ncbi.nlm.nih.gov/32565027/], specific target-ADR pairs have been filtered out whenever the ADR term fall in any of this particular SOC classes, which are not specific to body parts or underlying human biology, as pointed on the definitions of MedDRA’s guidelines (https://admin.new.meddra.org/sites/default/files/guidance/file/intguide_%2023_1_English.pdf)

• General disorders and administration site conditions
• Infections and Infestations
• Injury, poisoning and procedural complications
• Investigations
• Neoplasms benign, malignant and unspecified [incl cysts and polyps]
• Product issues
• Social circumstances
• Surgical and medical procedures
• Infections and infestations
• Psychiatric disorders

Output

Different files of output are obtained from the procedure:

• The databases cleaned versions, in the folder relationship_analysis_input_files. The latter are tsv files containing the access code of the different entries (if available) and the information of the drugs and the associated side effects
• The pairwise relationship file containing the computed p-value and q-value correction of the association target - side effect, called "qvalues_all_interactions"

Usage

1. Set up the postgresql server
2. Make sure that the connection in local works without a password (modify the postgresql file pg_hba as method "trust" for "local")
3. clone or download all the file in the repository and save it to a folder
4. Download the Athena dictionaries and save them into a folder called "athena" inside the repository folder
5. run the run_file.sh script (you need to have the permission to create/remove folder)

Remark

The Faers cleaning procedure has been based on the repository at https://github.com/ltscomputingllc/faersdbstats, the files has been updated to accept new FAERS AND MEDEFFECT data but the logic behind remains unchanged