Global acronyms are used in written text without their formal definitions. This makes it difficult to automatically interpret their sense as acronyms tend to be ambiguous. Supervised machine learning approaches to sense disambiguation require large training datasets. In clinical applications, large datasets are difficult to obtain due to patient privacy. Manual data annotation creates an additional bottleneck. This application automatically modifies PubMed abstracts to simulate global acronym usage and annotate their senses without the need for external sources or manual intervention. It can be used to create large datasets that in turn can be used to train supervised approaches to word sense disambiguation of biomedical acronyms.

To create an annotated corpus, start by providing a search query that describes a specific area of interest. The matching abstracts will be downloaded from PubMed automatically. Each abstract will be modified so that for each acronym it defines, either its full form or the acronym itself is used consistently throughout the abstract. This choice between the two is random. When an acronym is retained, its full form is used as its sense annotation. Once the corpus has been processed, it can be downloaded together with the sense inventory in a simple JSON format ready to be processed locally by other natural language processing applications.

The application is written using MEAN stack (http://meanjs.org/)

Demo is accessible on https://datainnovation.cardiff.ac.uk/acronyms/

The following will install the npm packages according to the configuration:

Run npm start for a dev frontend server and node server.js for backend server from separate consoles. Navigate to http://localhost:4203/. The app will automatically reload if you change any of the source files.

The following will build the production version of the applicaion:

The following will run the production version of the application:

To make use of all the APIs and mongoDB connection, you need to provide API Keys and mongoDB user settings in .env file in the source directory. The variables required:

• For Pubmed API: PUBMED_API_KEY
• For UMLS: UMLS_API_KEY
• MONGODB_USER: MongoDB username
• MONGODB_USER_PASSWORD: MongoDB password

Please refer to https://www.ncbi.nlm.nih.gov/home/develop/api/ in order to acquire PubMed database API key.

Please refer to https://documentation.uts.nlm.nih.gov/rest/authentication.html in order to acquire UMLS API key.

Additionally, if the intention is to run the application not locally, you need to register the application/domain at Google reCaptcha service (https://www.google.com/recaptcha/). Please provide an additional enviromental variable in that case: CAPTCHA_SITE_KEY. If that variable is not provided, a default value of the site key is used that works only for local environment.

• 30/06/2022:
  • Updated UMLS API (Reason: use of ticket granting tickets and service tickets was deprecated by UMLS)
  • Removed MongoDB connection (Reason: no need to store UMLS tickets as their use was deprecated by UMLS)

• Maxim Filimonov
• Irena Spasić

If you have any questions, feel free to contact the authors on spasici@cardiff.ac.uk or filimonovm@cardiff.ac.uk