Biomedical Search Evolves with novo|seek

Feb 03, 2009


      Bookmark and Share

The field of life science is always in flux. New ideas and theories are popping up by the minute, making it difficult for experts to stay on top of the newest, most timely information. Bioalma, an IT company based in Madrid, Spain specializing in the research and development of biomedical software, aims to simplify information discovery in the life science field with a new search engine, novo|seek.

Officially launching today, February 3, 2009, novo|seek is a free biomedical search engine designed specifically to help users extract information from published knowledge in biomedical literature. It is estimated that there are at least 18 million documents across more than 20,000 medical journals available to medical experts. With the abundance of information available, sifting through it all to find specific documents can become an overwhelming task. "We came up with this new product with a whole different philosophy. It is a search engine based on a text mining technology that will help people to get through a process that is time consuming and hard," says Ramon Alonso-Allende, director of marketing and business development for Bioalma.

Search engines such as Google Scholar and PubMed provide an overview of available literature, but according to Allende, they do not allow users to narrow their results. Instead, one must read through each returned result to see if the desired concepts are present. "Many search engines today find a bunch of results that aren’t relevant to what you want to do…With our tool, you can say which is the meaning you want to give to a certain term that you want to look for," explains Allende.

The ability to discern a user’s intended meaning comes from novo|seek’s text mining solution, which indexes biomedical literature and allows for recognition of biomedical terms. This indexing technology enables novo|seek to retrieve all documents where a term is mentioned, take into consideration synonyms, and reject documents where the desired term is used with a superfluous meaning. As Allende explains, "Our algorithm reads the text based on rules and training, and with the help of dictionaries, identifies where the term might be relevant to life science research." This way, if a user types in a certain term, not only do they get back results that include that term, they receive any document that includes biomedical synonyms of that term. For instance, if a users searches the terms "breast cancer" and "mitoxantrone," novo|seek will consider documents that include these exact terms, but also those containing "malignant neoplasm of the breast" and "novantrone", as terms that are synonyms of "breast cancer" and "mitoxantrone", respectively.

Novo|seek offers other features which aim to make searching easier for users. With this engine, users can highlight relevant biomedical concepts in the text, review information derived from documents on a single screen, search for an author and find research concepts based on the analysis of their research, and link to relevant external chemical and biological information. If a user searches for a certain disease, novo|seek will return documents where the queried disease is mentioned and also supply suggested information about drugs that cure that disease, genes that may cause it, or symptoms that are induced by the disease. Novo|seek also formulates a list of concepts relevant to the documents a user searches for, and when clicked, these concept titles are added to the user’s search along with the original query, narrowing returned results even more.

As of right now, novo|seek only catalogues abstracts from Medline, allowing users to a get a summary of search information without having to sort through entire documents. Eventually though, novo|seek aims to include full text documents, which Allende admits is a difficult undertaking. Other developments are in the works, as well. "We are planning to add U.S. grants. They provide lots of interesting information and give an idea on which direction life science research will go in the next few years," explains Allende.

Though adding more research is in the works, right now novo|seek is focusing on bringing the most relevant information to users in an easy way. "Before we started this project, we did research to see where we fit. We saw people liked Google because of the easy interface, so we tried to go that way," says Allende, "our end users don't want to bump into terms like text mining, they just want to search and get results, and don’t care how."

For novo|seek, having users worry about the correct way to search for desired information is not part of the plan. When using novo|seek, users should be able to focus more on their work, and less on struggling to find the answers they are looking for. "Our users are not really into being experts in searching, that was one of the things we wanted to avoid," says Allende. "With novo|seek, you don’t need to know you need to use brackets to search in certain fields, you can just focus on your work."

(www.novoseek.com)