Progress in the development of search engines and digital libraries in the last two decades has revolutionised access to biomedical literature. One only needs to recall that before the advent of PubMed and Google researchers had to mainly use library index cards to identify articles of interest. Besides giving instant access to papers based on keyword queries, search engine systems enable users to browse citation information, thus offering a much more efficient process for performing background research and literature reviews.
Nonetheless, keyword-based searches and citation analysis have their limitations. Examples of medical information needs that cannot be fulfilled directly by the current generation of search engines include "give me the list of all diseases and syndromes that are comorbid or are otherwise related to asthma" or "give me all pharmacological agents that have been used in the treatment of migraine".
Whereas the latter type of query is self-explanatory, the former is significant because asthma is a complex and multifactorial disease. Thus, the co-occurring conditions may share common aetiologies with one or more of its factors or may otherwise shed light on its causes, providing researchers and practitioners with additional avenues for understanding the disease and exploring new treatment options.
While there are online resources that can give partial answers to such questions (examples include WebMD, Medscape, The Mayo Clinic and Wikipedia), these are curated manually and therefore may not always be upto-date. The alternative is to supply broad queries to a search engine of choice, meticulously browsing and analysing the returned results to build the desired big picture. The results will be organised according to the search engine's global ranking criteria, which, from the user's perspective, are arbitrary with respect to the actual information need. This makes the browsing process unnecessarily time-consuming.
Memantic aims to tackle the challenges described above. Memantic captures relationships between medical concepts by mining biomedical literature and organises these relationships visually according to a well-known medical ontology. For example, a search for "Vitamin B12 deficiency" will yield a visual representation of all related diseases, symptoms and other medical entities that Memantic has discovered from the 25 million medical publications and abstracts mentioned above, as well as a number of medical encyclopaedias.
The user can explore a relationship of interest (such as the one between "Vitamin B12 deficiency" and "optic neuropathy", for instance) by clicking on it, which will bring up links to all the scientific texts that have been discovered to support that relationship. Furthermore, the user can select the desired type of related concepts — such as "diseases", "symptoms", "pharmacological agents", "physiological functions", and so on — and use it as a filter to make the visualisation even more concise. Finally, the related concepts can be semantically grouped into an expandable tree hierarchy to further reduce screen clutter and to let the user quickly navigate to the relevant area of interest.
We believe Memantic can save a considerable amount of time for researchers, medical students and practitioners who are investigating a particular condition, symptom or drug by giving a quick yet rich "mind map"-like overview of related medical concepts. We call our system a knowledge discovery engine because in contrast to traditional search systems it allows the user to quickly identify relationships that were previously unfamiliar to them. We achieve this in two ways:
Concisely organising related medical entities without duplication
Memantic first presents all medical terms related to the query concept and then groups publications by the presence of each such term in addition to the query itself. The hierarchical nature of this grouping allows the user to quickly establish previously unencountered relationships and to drill down into the hierarchy to only look at the papers concerning such relationships. Contrast this with the same search performed on Google, where the user normally gets a number of links, many of which have the same title; the user has to go through each link to see if it contains any novel information that is relevant to their query.
Keeping the index of relationships up-to-date
Memantic perpetually renews its index by continuously mining the biomedical literature, extracting new relationships and adding supporting publications to the ones already discovered. The key advantage of Memantic's user interface is that novel relationships become apparent to the user much quicker than on standard search engines. For example, Google may index a new research paper that exposes a previously unexplored connection between a particular drug and the disease that is being searched for by the user. However, Google may not assign that paper the sufficient weight for it to appear in the first few pages of the search results, thus making it invisible to the people searching for the disease who do not persevere in clicking past those initial pages.
Unlike a number of existing clinical decision support systems, Memantic does not suggest diagnoses based on symptoms and other observations. Instead, it simply exposes relationships between medical concepts, leaving it up to the healthcare practitioner to choose how to use any novel information they discover for the medical case at hand. One important aspect of our system is that we do not automatically infer causality for any such relationship, instead encouraging the user to read the supporting literature to establish any possible causal aspects for themselves. Our philosophy is that the doctor should be in charge at all times and should not be unduly influenced by automated suggestions. Instead of prescribing diagnoses or specific courses of action, Memantic offers a quick way to broaden the practitioner's horizons for a particular medical topic by highlighting relevant relationships that might otherwise be easily overlooked.