I have to use Pubmed several times every day and in most cases I have to switch to Google Scholar as I think that is really user-friendly and I can customize my search queries more easily. Although, I would love to do the same with Pubmed. Well, the Semantic MEDLINE Prototype which is a research and development project of the Cognitive Science Branch, Lister Hill National Center for Biomedical Communications, U.S. National Library of Medicine might solve my problems in the near future:
Semantic MEDLINE is a prototype Web application that summarizes MEDLINE citations returned by a PubMed search. Natural language processing is used to analyze salient content in titles and abstracts. This information is then presented in a graph that has links to the MEDLINE text processed.
Currently, the results from 35 PubMed searches (including a variety of disorders and drugs) are available to be processed. The 500 most recent citations (from the date of the search) are available for further processing by Semantic MEDLINE.
I just did a search for “Breast Cancer (clinicaltrials.gov), top 500 recruiting studies”:
Scienceroll.com readers know well I’m an admirer of WolframAlpha:
I use WolframAlpha because sometimes (if I know exactly what I want to find) it saves me plenty of time and clicks. If I want to calculate BMI, Google lists me several calculators. WolframAlpha calculates it itself. If I want to find information very fast about a clinical marker, Google gives me resources, WA gives me the best answer in one click. I also use it for ICD classification, as it’s more easily accessible than Wikipedia; for epidemiological data and other calculations.
To sum it up, I think WolframAlpha is for those who perfectly know what they want to find and want to save time and clicks. For other search queries, Google is still the best.
Now the Wolfram Alpha Team released a guide about how this unique search engine can be used for analyzing medical test-related data.
You can fine-tune the results even more by adding additional personal attributes. For example, entering “cholesterol tests age 65” filters the general population distribution to return only values from individuals 60–70 years old.
By adding more filters such as smoking status, diabetic status, pregnancy status, and other individual characteristics, you can find out more about how your test results compare to other populations covered by NHANES.
Just a short note about HealthMash, the health knowledge base, that is being developed by my friends at Polymeta.com and that was showcased at the MLA (Medical Library Association Annual Meeting and Exhibition) in Honolulu.
HealthMash™ combines sophisticated Web 2.0 universal search and discovery technology with Semantic Web Concepts in a simple yet highly informative user interface.
Here is an exclusive screenshot as this is the first time you can see the service from inside.
Attila Csordás, author of PIMM, published a screenshot on TwitPic about the new search engine, WolframAlpha:
It means the search engine understands what you want to find and gives you one specific answer, and not a list of possible answers. Huge difference, but that is what semantic search should be about.
WikiProfessional was officially launched some weeks ago. So I think it’s time to say a few words about it. WikiProfessional is a new kind of a database. It searches in several sources and helps us how to get the most valuable information.
Redundancy of the same facts and opinions within a myriad of web-pages has artificially inflated the size of the Internet. To get a million search results on a query without the ability to separate redundancy of the same information from the incremental knowledge expansions on that query concept is highly inefficient. Within the Concept Web, information is converted to streamlined knowledge where redundancy and newness of idea expansion are properly represented.
The sources it uses (yes, it searches in Wikipedia)
I gave it a try with cystic fibrosis. Here is what I got:
A concept tree with the articles that should be mentioned
A proper definition, functional information, etc.
We still need time to get used to this system but I’ m pretty sure it can be better and more user-friendly than Pubmed itself.
Reviews and debates about it:
I’ve been playing with Powerset for a while. It seems to be a service that can take us to the world of semantic web or web 3.0. It uses Wikipedia and Freebase as resources. The main idea is to ask questions instead of search for terms. Let’s give it a try.
If you make a search for “Who discovered penicillin” in Google, you will see this (Alexander Fleming Discovers Penicillin) and many more similar articles. Even if we know the truth is different.
If you ask the same question in Powerset, you get this:
It’s a bit more accurate, isn’t it?
Try it and let me know if you find something interesting. And don’t forget to check out my Personalized Medical Search!
Here are some more examples.
The most famous article ever written about web 2.0 and medicine belongs to Dean Giustini (UBC biomedical branch librarian) who now made an other big step in this special field. Check out the new article (Web 3.0 and medicine) on British Medical Journal online.
A great honor for me is that Scienceroll has been used as a reference:
Social software enthusiasts may well find that the new web will be fertile ground for the creation of knowledge. Although already popular, wikis may well serve as platforms for the exploration of web 3.0. One innovative wiki—Wikiproteins—is already using semantic technologies. In contrast to other wikis, Wikiproteins imports data mined from several of the world’s leading biomedical databases, such as PubMed, UniProt, and the National Library of Medicine. Its integrated entries are a useful combination of genetic information and scientific literature. Notably, the confluence of databases in Wikiproteins yields more than two million factual associations for data mining and over five billion associated pairs.9
9.) Mesko B. Web 3.0 and medicine. ScienceRoll blog. 2007. http://scienceroll.com/2007/04/06/web-30-and-medicine/.
If you want to read more on the subject: