Author ORCID Identifier
Date of Award
Department or Program
Quantitative Biomedical Sciences
Margaret E Ackerman
The development of a vaccine for Human Immunodeficiency Virus type 1 (HIV-1) is a crucial step in preventing the global spread of AIDS. To ensure the progress and effectiveness of this vaccine, it is essential to establish efficient biotechnology platforms and data mining methods. These methods would help identify immune characteristics that distinguish individuals with varying levels of vaccine-induced protection and determine the underlying factors that contribute to protection against HIV acquisition. While previous studies have focused on identifying immune markers associated with infection outcomes among vaccinated patients, it is important to acknowledge the limitations of traditional case-control analytical protocols due to the difficulty in obtaining comprehensive exposure information. However, the concept of transductive Positive-Unlabeled (PU) Learning methods can be adapted to address this challenge. By inferring the level of protection based on a limited number of samples labeled as "infected" or positive, this type of approaches demonstrated potentials for further investigation. The primary objective of this study is to develop and evaluate an analysis pipeline based on PU learning. This pipeline aims to infer patients' levels of vaccine-mediated protection using high-dimensional immunological profiles and identify potential immune markers for future vaccine research. The study includes synthetic datasets under various hypothetical scenarios, real-world immunogenicity data with known group labels, and clinical endpoints to demonstrate the effectiveness and sensitivity of our selected PU-based methods. Furthermore, we have developed a methodology that combines multiple state-of-the-art PU learning and statistical methods. This methodology serves as a validation approach for cases where explicit negative examples are not available. By applying this pipeline to a real-world dataset and obtaining robust results, our work provides a blueprint for using PU learning in the identification of correlates of protection in contemporary immunological research. We anticipate that continued advancements in vaccinology, data measurement technology, data mining techniques, and clinical trial design will further enhance our understanding of immune responses and contribute to the development of effective vaccines.
Xu, S., Carpenter, M.C., Spreng, R.L. et al. Impact of adjuvants on the biophysical and functional characteristics of HIV vaccine-elicited antibodies in humans. npj Vaccines 7, 90 (2022). https://doi.org/10.1038/s41541-022-00514-9
Natarajan, H., Xu, S., Crowley, A.R. et al. Antibody attributes that predict the neutralization and effector function of polyclonal responses to SARS-CoV-2. BMC Immunol 23, 7 (2022). https://doi.org/10.1186/s12865-022-00480-w
Xu, Shiwei, "PREDICTING HIV VACCINE-MEDIATED PROTECTION LEVEL TO IDENTIFY IMMUNE CORRELATES USING POSITIVE UNLABELED LEARNING" (2023). Dartmouth College Ph.D Dissertations. 186.
Available for download on Wednesday, June 18, 2025