Translator Red Knowledge (TReK)

Columbia University Health Sciences, New York NY

Investigators

Linked publications, trials & patents

Abstract

Rationale: The goals of this application will be most effectively and efficiently met with a multiple PI plan. Drs. Weng, Dumontier, and Ta are recognized leaders in their fields and are excellent choices to ensure that the goals of the project are carried out in the most effective manner possible. Governance and Organization Structure: Drs. Weng, Dumontier, and Ta will share leadership equally for the project and each will take responsibility for specific organizational functions. Specifically, Dr. Weng will serve as the Contact PI. While all portions of the project will have input from all PIs, primary responsibilities will be split as follows: Drs. Weng and Ta will jointly oversee the following tasks: 1) all enhancements of the COHD data analysis pipeline and COHD API, including temporal analyses, processing clinical notes, and COHD replication studies; 2) development of representation learning methods for patient vector embedding and concept vector embedding, including downstream analyses like disease subtyping; and 3) creating the RCT PICO Knowledge Provider from output from the PICO NLP Parser. Dr. Dumontier will oversee the following tasks: 1) development and deployment of Data2Services; 2) development of the SPARQL Reasoner API interface to expose BioLink-compliant RDF as a Reasoner API; 3) development of user interfaces for querying and navigating knowledge graphs; and 4) management of the DSRI OpenShift cluster. Together, Drs. Weng, Dumontier, and Ta will oversee the preparation of the annual progress reports and other communications with the funding organization. PIs will work together to manage policy issues and to communicate with deans and other administrators to promote the success of the program. Communication Plan: While the PIs interact frequently in informal ways (at seminars, in other joint activities, and as colleagues and friends), they will also institute formal monthly web conferencing meetings to discuss study design and procedures, review milestones, interpret outcomes, and all administrative responsibilities. The PIs also plan to communicate frequently by e-mail, phone, and in person as issues arise that must be addressed promptly and also to review the success of initiatives and make programmatic adjustments as needed. Conflict Resolution: Conflicts will be resolved first through joint meetings between Drs. Weng, Dumontier and Ta. If a potential conflict develops, the appropriate departmental administrators representing the PIs shall meet and attempt in good faith to settle any dispute, claim or controversy arising out of or relating to the interpretation, performance or breach of this disagreement. However, if the departmental administrators fail to resolve the disagreement within thirty business days, then such disagreement shall be referred for resolution to a designated senior executive (e.g., deans) of the parties who has the authority to settle the disagreement but who is not directly involved in the disagreement. Intellectual Property: Since the NCATS Biomedical Data Translator program is principled on developing open-source tools, all products will be released with open-source licenses. Distribution of Resources: To the extent possible, the distribution of resources will be evenly divided between the two institutions (Columbia and Maastricht University) in line with the methods indicated in the application. Change in PI Location: If any PI moves to another institution, all possible attempts will be made to transfer the relevant portion of the grant to the new institution. If the new location is unable to accept or support the research project, then an alternative PI will be chosen to pursue that part of the project, if possible.