Predictive models on temporal data in the form of observations in a sequence or series have helped make headway in utilizing large sets of clinical data collected from health systems in the electronic medical record and billing systems to predict outcomes, diagnoses and help make quicker decisions for health of populations with precision, treating the person and their disease based on genomic characteristics or predictive modeling. This concept of precision medicine is a heavy focus in research on diseases like cancer, and it is grounded in the individual’s cancer genomics which could have completely different molecular characteristics case-to-case and person-to-person. In fact, the tumor itself has different genomic characteristics when compared with other cells in the body. This proposal intends to look predictive models based on temporally based observational health data that is sequential in nature from clinical data sets to annotate the complex biology collected from cancer patients – specifically children suffering from highly lethal rare high-grade gliomas. This research focuses on the intersection two national longitudinal health data collection projects: The Children’s Brain Tumor Tissue Consortium (CBTTC) and the PEDSnet Clinical Data Research Network (CDRN) with the intention of harmonizing the two national projects as they grow to aid in the human annotation of biologically based resources with large scale automated health data networks. The research proposed herein contributes to the coherence broadly defined technical area of cross-platform workflow architectures to facility resource sharing and reuse in biomedical research. Specifically, the methods in this proposal use two common observational research methods utilized by two national organizations to come up with ways of doing large scale data and system integration for rare and deadly diseases by opening this data to more minds outside of purely the biomedical domain.