MedScan leverages hand-crafted dictionaries of protein and chemical names, and precise text parsing and pattern matching algorithms specifically developed for biological text analysis. It extracts functional relationships between proteins, cell processes, and small molecules, recognizes types of regulatory mechanisms involved and the effects of regulation, and can be customized to extract other types of information.
"The majority of pharmaceutical and biotechnology companies recognize the need for an in-house information extraction technology that would help them to process literature, patent applications and internal text documents in order to create and automatically maintain internal functional relationship databases, amenable to computational analysis," said Nikolai Daraselia, Director of Research at Ariadne Genomics. "Data extracted with MedScan and manually curated data are normally of compatible quality, but the efficiency is incomparable, and it is easy to verify and replicate the results of MedScan work."
The new version of MedScan adds flexibility on the level of entity recognition and information extraction. It has become simple to customize dictionaries, and extract new types of relationships by adding or editing linguistic patterns. MedScan users can create information extraction pipelines with different output, set up regular automatic processing of upcoming content for selected sources, and integrate MedScan with other bioinformatics tools for further analysis of extracted information.
For more information about MedScan, or for a web-based trial, visit Ariadne Genomics website www.ariadnegenomics.com/products/med