Searching for a search engine

One of the things I love most about being a scientist is constantly learning. Normally that means regularly searching for new papers published in my field of research, either by entering the term/gene/topic in PubMed or by scanning the tables of contents of a handful of journals I follow. However, every once in a while an experiment spits out an unexpected result that changes the course of the whole project, linking my research to something I know nothing about.

ASCB Navigator provides interactive charts of researchers
and their collaborators. Here’s a map of three collaborators
(Pierre Wong, David Borchelt and Don Cleveland)
and their colleagues, who have published 16 or
more papers together.One of the things I love most about being a scientist is constantly learning. Normally that means regularly searching for new papers published in my field of research, either by entering the term/gene/topic in PubMed or by scanning the tables of contents of a handful of journals I follow. However, every once in a while an experiment spits out an unexpected result that changes the course of the whole project, linking my research to something I know nothing about. The last time this happened I typed a single word in PubMed and got more than 100,000 hits! I knew I needed to come up with a better approach to this and decided to take a look at other options by, yeah that’s right, searching for a search engine.

After trying several search engines I realized there are two main ways engines search for articles: 1) by searching for keywords and spitting out any article containing the specific word(s) or 2) by arranging articles into a hierarchical tree according to their topic and showing all articles within the group related to your search word. In general, I prefer type 1 for quickly finding an article when I know exactly what I am looking for, but find type 2 engines more useful when I am just exploring a new topic. Here’s a list of my top five searchers:

1. Quertle. Pretty much like PubMed, Quertle searches mainly by keywords, but I prefer Quertle over PubMed for several reasons: It displays results in two categories (Focused Results and Broader Results) according to relevance. Quertle allows you to search using verbs (i.e., increase, decrease, regulate) that modify the keyword but are not treated as keywords themselves. Plus, Quertle searches journal articles and other scientific sources, like grant applications.

2. gopubmed is a structured search engine where articles are sorted by topics. Searching with gopubmed begins like any other search, by typing a keyword, but unlike other engines the results in gopubmed are organized in a hierarchical tree that can be easily accessed through a navigation bar. This navigation bar breaks up all results into common “concepts” or groups of papers with a similar topic. This handy feature makes gopubmed the search engine of choice when I am looking to learn about a new topic, as I can use the “concepts” to guide my search.

3. PubMed is like your favorite mother’s recipe: you know it well and it will fill up your craving, as long as you know what you are craving. Plus, with a click of your mouse you can turn your search into a keyword search, a gene search, a protein search, you name it. Problem is, it is still the same ol’ good-but-not-great search engine you’ve come to love and hate.

4. Google Scholar has the power of a Google search with a touch of science. It displays not only the paper title but the authors and citations. It also allows you to create a profile for your papers, keep track of who is citing you, and makes suggestions of papers to read based on your published articles. Pretty neat, except that it is still a Google search and therefore filters are not easily available to help narrow your search.

5. Scirus made the cut because it is by far the most comprehensive science search engine I have encountered. Scirus searches not only indexed journals but also patents, scientist websites, institutional repositories, tutorials, and even the Encyclopedia Britannica. This comprehensiveness can make Scirus your best friend or worst enemy. For example, a search using the term “mitosis” gets just over 60,000 hits in PubMed. In Scirus the same search racks up more than 600,000 hits! So, I recommend using Scirus only for hard-to-find narrow searches or when you want to spend a whole week sorting through the results.

Although I like having multiple options for my searches, sometimes it is difficult to decide where to start. The good news is that a new generation of search engines is being born that can steer us in the right direction. For example, the ASCB has just launched a new members-only, three-in-one search engine called the ASCB Navigator. It is not exactly a search engine that one would use for literature searches, rather a science intelligence tool to understand who the key players are in one’s favorite field, and news of what’s happening in the science business. It has a BioExpert Navigator that identifies experts in your field of interest and links them to their collaborators in an interactive chart. It also has a Clinical Trials Navigator that allows easy searching of clinical trials, a BioNews Navigator to search within syndicated news sources, and a Research Navigator to search grants and patents. The ASCB Navigator has only been live for less than two weeks and so I still need to kick the tires since it uses an approach different from what I am used to. But so far I am loving the ability to search multiple information sources in one single engine.

Do you have a favorite search engine? Please do share the information. I am always looking for better ways to navigate the sea of scientific publications.