Automation of Precision Medicine for the treatment of AML through development of CAR NK Cells

Patrick Kim

Mentor: Dr. Sijung Yun, Predictiv Care and Johns Hopkins University.

Date/Time: August 23rd, 2024 at 2:20pm.

Abstract: Chimeric antigen receptor engineered natural killer cells have yielded promising results in the treatment for patients afflicted with acute myeloid lymphoma. Tumorous AML cells exhibit antigen specificity in which CARs can be designed to target. Genome analysis of patients with AML can be analyzed for the development of catered precision medicine to target AML with specificity regarding the antigen exhibited in their tumors. This project therefore focused on the automation of creating a pipeline that would focus on developing CARs that could be engineered with better efficacy and higher rates of affinity for binding to AML tumors.

In August, I began fine-tuning the newly released SAM2 model. This work aims to apply masks to all particles within a micrograph, a task that may not be fully realized by the time of this presentation. The work completed thus far has laid a foundation for future improvements in cryo-EM data collection using AI-driven methods.

The results of the predicted modeling and docking software showed promising results that implied precision medicine catered to specific patient mutations in AML was possible. The automation of developing sequences and models for the specific CAR was successful; however, certain parts were not easily automated. Sequence analysis for family and domains for the proteins and docking software was not easily achieved and still required manual inputs.