Fujitsu and Kyoto University Start R&D for AI that Infers Clinical Interpretations of Genome and Gene Polymorphism

Tokyo, October 06, 2016

Fujitsu Limited and Kyoto University today announced that their R&D project to build a clinical genome knowledge and information platform will be adopted into the Project to Create an Integrated Database for Clinical Genome Information under the Japan Agency for Medical Research and Development (AMED). Under the project, led by Professor Yasushi Okuno of the Kyoto University Graduate School of Medicine, Fujitsu will lead the drive to develop AI technology in support of doctors and other healthcare professionals' efforts to provide clinical interpretation. Plans are for the project to continue for a period of five years (until 2020).

This research project has as its goal the realization and spread of genomic medicine, which is the use of genetic information in the diagnosis and treatment of disease. It aims to build an integrated clinical genome database that provides highly reliable and accurate clinical interpretation to the public, integrating databases of clinical information and genomic information for each type of condition, such as cancer, along with all sorts of public databases, which AMED is already working on creating.

Background and Goals

In bringing about the realization of genomic medicine, in which genetic information is used in the diagnosis and treatment of disease, an issue that has arisen is the building of an Integrated Database for Clinical Genome Information, as a knowledge platform for enabling the use of this information more broadly in clinical settings. In the US, for example, as a framework for aggregating gene polymorphism information related to diseases and associated clinical interpretation, databases that can be freely accessed by specialist organizations are being built and announced as quickly as possible.

There have been issues, however, in the reliability and accuracy of the data that depends on the clinical facilities and organizations which provide the clinical interpretation. For this reason there are ongoing efforts to improve reliability and accuracy, such as by systematizing a mechanism for evaluation by experts. In Japan, as well, the construction of database platforms is moving forward, and there is a demand for initiatives to standardize and systematize data aggregation and the provision of clinical interpretation.

Project Efforts

Working together with Kyoto University, Fujitsu is now developing a curation system to formulize and put algorithms into the process of aggregating data, which was previously done by experts by hand, and to support the aggregation of data by experts. With this curation system, Fujitsu aims to support the provision of highly accurate and rapid clinical interpretation by developing unique machine learning and AI technology.

In the development of this curation system, Fujitsu is working on R&D for the below technologies to infer clinical interpretations for gene polymorphisms for which the relationship to disease is not currently clear.

• Fujitsu is building a knowledge base that aggregates academic literature and public databases in the medical field, using the Fujitsu Laboratories-developed Linked Open Data (LOD) utilization　platform⁽¹⁾.
• Using this knowledge base, Fujitsu aims to build a system that uses the unique machine learning technology that makes up Human Centric AI Zinrai, Fujitsu's AI technology, to infer clinical interpretations, and to supply the evidence for that inference and candidate treatments.

Fujitsu plans to develop these technologies, and build a business around them as AI technologies that support R&D at companies and universities, not only in healthcare but in a variety of fields.