Ramy Arnaout, a clinical pathologist at Beth Israel Deaconess Medical Center in Boston, believed he and his colleagues were wasting precious time performing Web searches for scientific journal articles. Time would be better spent curing diseases and treating patients. So he developed an application at first to use on his own computer to speed up his online queries for medical texts, but his efforts have evolved into a Web-based search tool for life sciences research papers and a company to market the technology, Pubget.
Cambridge, MA-based Pubget was founded in 2007 and recently disclosed that 50 research institutions have adopted its search service. The six-person startup’s search engine is designed to retrieve full-text research papers on PDF files with a single query. This is intended to be an improvement to previous science research paper search tools such as the National Institutes of Health’s PubMed and Google Scholar, both of which often require people to click through multiple links to get to a full-text version of a journal article. The company boasts that its search engine can retrieve some 20 million articles from 20,000 life sciences journals.
I gave Pubget a whirl last week, having had some troubles of my own finding life sciences texts with PubMed. I started my search with a toughie, “natural killer T cells,” and found that Pubget found 23,371 documents, with the first 100 results listed on the left hand column of the webpage. (Ryan Jones, president of Pubget, gave me a quick primer on how to use the startup’s search engine.) As Jones had told me, Pubget only displays the full-text PDFs of research papers if they are either free or your institution has a subscription to the journal that publishes them. Of the few NKT cell articles I clicked on, the most I was able to get was an abstract. Unfortunately, only about a third of the articles Pubget is able to retrieve are free. (Our tax dollars pay for lots of the research published in subscription-only journals, but that’s a debate for a different time I suppose.)
Yet there are a few innovations that Pubget appears to bring to the table. First of all, other search engines that I’ve used are not able to retrieve a full-text PDF file—the kind that make for easy, readable printouts—in a single search. Second, Jones tells me that Pubget is programmed to know whether your institution has a subscription to a journal that publishes the article you want, using single sign-on technology the firm developed to enable full articles to be displayed without asking for additional passwords and such. Even researchers at Harvard, which has subscriptions to many journals, can’t easily retrieve articles online with previous search engines because of this need to keep re-entering passwords, he says.
Pubget’s business model is to provide the search service for free, and to make money by selling ads and other marketing products to the life sciences companies. If a search finds an article that, say, covers research involving a high-powered microscope, microscope manufacturers are willing to pay a premium to display their ads to that researcher, and pay for lists of potential customers from Pubget’s search data. But the company doesn’t reveal data on individual users and their searches, Jones says.
Arnaout, who serves as CEO of Pubget and remains a practicing physician, filled me in on the journey he took to found the startup. Actually, he tells me, the journey didn’t begin with a business venture in mind. It was prompted by the fact that he couldn’t load the search app he developed for his Mac onto his colleagues’ PCs. He eventually sought some advice from his friend, Sami Shalabi, a former MIT classmate. Shalabi—who has worked for Google since the Internet giant acquired his mobile technology startup Zingku in 2007—advised Arnaout to consult Ian Connor, who was then a software developer for IBM. Connor, who is now chief technology officer of Pubget, assessed the problem and told Arnaout that he solution was to make the search technology Web-based rather than developing client software.
Connor also couldn’t understand how scientists could tolerate such cumbersome searches for science documents. “[Connor] said that if this were computer science this problem it would have been solved 20 years ago,” Arnaout says.
Connor and Arnaout spent lots of late nights and weekends for the better part of a year to develop a prototype of the search engine. Connor eventually left his job at IBM to dedicate himself to Pubget. Jones later joined Pubget via Microsoft’s enterprise search group. The startup is angel-funded by doctors and wealthy individuals.
Pubget’s stated mission is to save scientists time so they can find cures for diseases. Arnaout, who has a PhD in mathematical biology from Oxford University, says he did a calculation and determined that a faster search tool for scientific journal articles could save researchers worldwide about a half a billion minutes per year. “You can check my math,” Arnaout says, “but that’s like a small research team working full time since [the birth of] Christ.”
By posting a comment, you agree to our terms and conditions.