Currently, we index more than 22 million Life Sciences research abstracts at SciCurve. With the help of natural language processing and data mining, we are able to provide you with helpful information. Today, we would like to show you how to determine the most investigated factors for prostate cancer development, using SciCurve map view. The map view takes the top 1000 articles for given topic and places them on the map, generating clusters of articles with semantically similar focus.
Lets start with query: “increases risk of” prostate cancer. We will get a map like this one:
The age and relevance of published articles is visualized as color and size of nods representing them. As you can see, the map contains a lot of topics regarding cancer therapy. Also the deprivation cluster is mainly focused on “androgen deprivation therapy”. You can find this when you open the link and hover over the bubble in the map like this:
This is somehow useful, but we can improve our query. We should exclude articles about therapy using query: “increases risk of” prostate cancer -therapy. We will get a map like this:
Asa result we can observe main factors considered to affect the development of prostate cancer (you can evaluate them further by hovering over papers and checking titles). Here is the list:
- CAG repeat in androgen receptor
- BRCA2 mutations
- Vitamin D
- Fatty acid
- Selenium (not in the map but its dominant in one cluster; our technology is not perfect yet)
- Several polymorphisms
- And of course, family history
Below the graph, we can view the table of newest and most relevant research papers published for given factors:
We believe, that tools for easy interpretation of scientific results could reduce the occurrence of uninformed opinions in public debate. Simplified, yet standardized, communication of relevant scientific knowledge to the world, is the core value of SciCurve.com. We are working on simplifying the search and the use of our tools. Our target is to bring global research closer to general public and relevant meta-data to professionals.
If you would like to engage in any kind of cooperation with SciCurve.com, please contact us at: firstname.lastname@example.org
We value your support greatly.