Congratulations to the 2024 ATIP Recipients
Funded by a Clinical and Translational Science Award (CTSA) from the NIH National Center for Advancing Translational Sciences (NCATS), Accelerated Translational Incubator Pilot (ATIP) Program Grants seek to support projects that focus on the science of science and ultimately promote NCATS’ mission to accelerate translational research by exploring more efficient strategies for overcoming barriers to investigational studies. Grants are awarded annually and the theme for this years’ funding cycle was improving the process of data integration. The research proposals selected for support included 2 projects at Johns Hopkins and 2 projects from the ICTR’s CTSA partner, the University of Maryland Baltimore ICTR. Each awardee receives a 12-month, $50,000 grant.
2024 ATIP Recipients
Harrison Bai, MD
“Enhancing Large Vision-Language Model Efficacy through Representative Slice Selection from 3D Medical Imaging Data”
This project seeks to overcome limitations encountered in the ability of vision-language models (VLM) to transition from analyzing 2D to 3D medical images. The primary objective of this work will be to adapt large VLMs to interpret and analyze 3D medical imaging effectively. This will be achieved by developing a multimodal dataset with annotated 3D MRI brain scans; implementing as well as evaluating an unsupervised method for selecting representative 2D slices from 3D scans; and creating, a benchmarking system for assessing the clinical efficacy of VLMs.
Casey Rebholz, PhD, MPH
“Plasma Protein Biomarkers of Healthy Dietary Patterns, Chronic Kidney Disease Progression, and All-Cause Mortality”
This research proposal is designed to integrate untargeted proteomics data with dietary data, clinical characteristics, and health outcomes in the Chronic Renal Insufficiency Cohort (CRIC) study. With a focus on making the rich proteomics data available in the CRIC study accessible to more investigators, the major goals of this work include quantifying the quality of the broad and untargeted proteomics dataset; developing and implementing a processing pipeline for the proteomics data; and finally integrating the datasets with the clinical data in preparation for analysis.
Chikiang Chen, PhD, MS
University of Maryland
“A New Framework for Information Integration and Its Application to Study Risk Factors Associated with Alzheimer’s Disease Onset”
A statistically informed data integration framework will be developed to allow for the inclusion of information from diverse and large cohorts that can accommodate heterogeneity in the data (i.e., different outcome measures; predictor sets) and does not require accessing raw data from external sources that may include sensitive participant information.
Graeme Woodworth, MD
University of Maryland
“Predicting Bioeffect Outcomes of Microbubble-assisted Focused Ultrasound for Patients with Glioblastoma through Advanced Data Science”
The aim of this proposal will be to develop a feasible approach for achieving multidimensional data integration that will help overcome current constraints especially when conducting research studies in the emerging field of focused ultrasound (FUS). This proof-of-concept work will be accomplished by bringing together data from different databases containing complex and disparate domains of information, including patient attributes, technical parameters of a FUS procedure, plasma analyte levels, and neuroimaging characteristics, into an integrated dataset.
