As biomedical researchers, we rely on peer-reviewed journal articles to improve our knowledge and stay abreast of current information, techniques, and ideas. Journal articles often contain interrelated concepts of biological systems including biochemistry, immunology, molecular biology, genetics, diseases, and drug therapies, making it difficult for any single reader to know details of all the content. Therefore, we often resort to doing keyword web searches while reading to learn specific details about them. This is an inefficient and relatively time-consuming approach that may not always yield the desired results due to search engine rankings and a large number of search results returned.
To address this issue, scientists at the UTHSC Center for Biomedical Informatics and Department of Pharmacy, spearheaded by Panduka Nagahawatte’s (pictured right) lab, developed Genotation that enhances researchers’ reading experience with supplementary content. UTHSC experts Ethan Willis, Mark Sakauye, Rony Jose, MPH, Hao Chen, PhD, and Robert Davis, MD, MPH, played key roles in this initiative. With Genotation, a user can visually interact with supplements while reading the article as well as share a summarized graphical representation of the supplements with colleagues. The experience begins by uploading an article on www.genotation.org, where the Artificial Intelligence (AI) module sets on a quest to search and provide a wide array of information.
We now invite UTHSC researchers to play a key role in this initiative by becoming a Genotation user for free. Although designed to be intuitive, there is a video tutorial available in the help section, if needed. We are continuously improving Genotation to address three major challenges: (1) accurately identifying the subset of terms from an article to be linked to external knowledge, (2) synthesizing a centralized knowledge base containing information from multiple databases, and (3) present the supplementary knowledge intuitively to enhance the reader experience. We urge you to use this service and provide us feedback through either the surveys in the help section or by emailing panduka@uthsc.edu regarding the improvement in user experience, additional data sources and species of interest, other applicable scenarios, and the AI module that searches for genomic terms. Furthermore, we solicit datasets that relates biomedical knowledge to gene symbols, from scientific good Samaritans, to further enhance the knowledge base of Genotation.
Genotation currently contains information regarding 59,905 gene symbols, 5633 drug – gene relationships, 5981 gene-disease relationships, and 713 pathways. We are currently expanding the knowledge base with information on mouse and rat. The platform currently accepts documents in Portable Document Format (PDF) or Hyper Text Markup Language (HTML). We are currently assessing the possibility of expanding to Word, Excel, and PowerPoint documents.
Readers can upload PDF or HTML documents, at which point the AI module searches through the document to identify genetic terms that could be linked to supplementary information. Subsequently, each identified term is queried against the knowledge base to gather descriptive, functional, clinical, and pharmacogenomics information as supplements. These terms are categorized into genes, diseases, and drugs. The application displays the article in its original format, presenting the user with an unhindered view for reading. Supplements are stored inside an interactive, expandable, and searchable menu accessible to the user while reading the document.
The service is freely available on www.genotation.org. We look forward to your feedback.
-Panduka Nagahawatte, MS Staff Scientist, UTHSC Center for Biomedical Informatics