Andorf Lab Research Projects
Extension of clinical data structured into ImmPort
Public data sharing promotes transparency, reproducibility, and secondary analysis. However, clinical trial data is often shared in unstructured formats, making it difficult to reuse effectively. In this project, we work with NIH/NIAID’s Immunology Database and Analysis Portal (ImmPort), an extensive, open-access repository containing clinical and mechanistic data with an immunological focus. The overarching goal of this project is to develop computational tools that help data contributors share their clinical trial data in a more structured format, enhancing the accessibility and usability of the data for secondary research. (Funding: NIH/NIAID R03AI185587)
Secondary analyses of existing clinical study data
Sharing research data provides numerous opportunities for additional analyses beyond the initial studies. In several projects, we use publicly and institutionally available individual-level data from clinical studies to perform secondary analyses. One example is our recent work identifying risk subgroups among infants from the paradigm-shifting Learning Early About Peanut Allergy (LEAP) trial. This analysis focused on the probability of developing peanut allergy and estimated intervention effects of early-childhood peanut introduction. The individual participant data used for this study are publicly available through the Immune Tolerance Network (ITN) TrialShare and ImmPort platforms.
Data-driven modeling of early childhood food allergy
Detecting when a milk or egg allergy has been naturally outgrown in young children is paramount for enabling the earliest safe reincorporation of these foods into the child’s diet. This is important because prolonged milk and egg avoidance is associated with growth and nutritional deficiencies. In a project funded by the Gerber Foundation, we utilize machine learning to develop predictive models that help identify when egg and milk allergies resolve. Our overarching goal is to enhance precision for the diagnosis and resolution of these food allergies to reduce their burden on children and their families. By doing so, we aim to promote equitable care and better health outcomes. (Funding: The Gerber Foundation)
Computational approaches for cytometry data
Computational methods are needed to analyze high-parameter flow and mass cytometry in an objective and efficient manner. Our work includes the development of tools that address different aspects in cytometry analysis pipelines. Recent examples include CytoPheno, an automated tool that assigns marker definitions and cell type names to unidentified cell clusters, and cytoFlagR, a framework to objectively assess batch effects in high-parameter cytometry data.
