Chemotherapeutic co-medications for multi-drug resistant cancer are used to increase the efficacy of chemothereputics. Our core objective was to determine a method to identify properties of successful co-medications. Co-medication research is relatively new, so little data exists to identify good candidates from a cohort of existing drugs. Even with some eliminations/reductions, over 5 million potential co-medications need evaluation. In early research, our collaborators at the SMU Center for Drug, Discovery, Design and Delivery purchased 71 medications and tested them as potential co-medications, with 23 showing some success and the remaining 48 not. Each drug has 30 properties specifically relating to the compound and over 100 additional properties related to how the compound performed in vitro. For the initial problem, the researchers seek a way to determine commonalities in the 23 successful drugs in order to create a system to identify new co-medication candidates.
To tackle this problem, the property analysis was broken into microtask that were compatiable for Human Compuation Gaming (HCG). This task strcture was modeled after k-means clustering and allowed players to manual adjust centroids for two clusters on 2 dimensions (co-medication properites) at a time. The image below showed the proof of concept interface used to develop the interaction model.
The above interface was then converted into the Twin Stick Shooter HCG Wiley Wizard, developed by BALANCED Media | Technology, and deployed on the HEWMEN platform. Below is a trailer of the gameplay from Wiley Wizard
The initial results obtained from Wiley Wizard allowed us to evaluate the methodology for integration and evalution of human computation for both custom and commercial video games. The results were published in the proceedins from the 2018 IEEE 6th International Conference on Serious Games and Applications for Health (SeGAH)