Creating Investigator Networks from PubMed Data
Author Information
Author(s): Yu Wei, Yesupriya Ajay, Wulf Anja, Qu Junfeng, Gwinn Marta, Khoury Muin J
Primary Institution: National Office of Public Health Genomics, Centers for Disease Control and Prevention
Hypothesis
Can we automatically generate detailed investigator profiles and networks using PubMed abstracts?
Conclusion
The study successfully developed a web-based prototype that creates domain-specific investigator networks by accurately generating detailed investigator profiles from PubMed abstracts.
Supporting Evidence
- The parsing strategy extracted country information from 92.1% of affiliation strings in a random sample of PubMed records.
- The method identified 70-90% of investigators in three different human genetics fields.
- The prototype system successfully identified 9 of 10 genetics investigators within the PREBIC network.
Takeaway
This study shows how we can use information from research articles to find and connect scientists working on similar topics.
Methodology
The study used 20,000 randomly selected PubMed abstracts to develop a parsing tool that extracts affiliation and email information.
Potential Biases
Ambiguity in author names may lead to misidentification of investigators.
Limitations
The method can only generate profiles for first authors, and inconsistencies in institution names limit full automation.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website