Market Overview:
Hypothesis generation is becoming a crucial technique that allows researchers to make connects between concepts. Typically, the systems operate on domain-specific fractions of public medical data. There are many unknown connections in the biomedical sciences. In order to fill the gaps in existing biomedical knowledge, Clemson University researchers have developed MOLIERE, which utilizes information from over 24.5 million historic documents including scientific papers, keywords, genes, proteins, diseases and diagnoses. MOLIERE will propose a possible connection between biomedical objects that are not known to be related to one another.
Application Stage of Development
Biomedical Discovery, Data Analysis Prototype
Advantages
- Has a different algorithmic procedure from existing technology.
- Has massive validation results.
- Potentially can be used in the drug discovery pipeline.
Technical Summary
This technology has a multi-modal and multi-relational network of biomedical objects extracted from several heterogeneous datasets from the National Center for Biotechnology Information. MOLIERE models hypotheses using Latent Dirichlet Allocation applied on abstracts found near shortest paths discovered within the network. Effectiveness is demonstrated by performing hypothesis generation on historical data.
View printable PDF version of this technology
____________________________________________________________________________________________
Inventor: Ilya Safro
Patent Type: N/A
Serial Number: N/A
CURF Ref No: 2020-028