Lewis Frey, PhD develops novel algorithms and information systems for the purpose of discovery and data integration applicable to precision medicine. He has extensive experience working with virtual machine deployments in networks and in big data technology deployed within Veterans Affairs (VA). In his research on Clinical Personalized Pragmatic Predictions of Outcomes (C3PO) he brings both of these areas together to create a quickly deployable system for building networks and sample size using the C3PO Big Data System he built for the VA. In addition to applying novel machine learning to medical data, his information systems approach combines the accumulated wealth of knowledge that exists in medical record systems with the vast amounts of molecular data being captured with high-throughput measurement technologies. The following are active areas of his research:
Investigation of the utility of novel similarity measures to conduct predictive analytics
---The goal is to improve outcomes through utilizing similar patients to inform treatment approaches. Dr. Frey is interested in applying this approach to characterization of medical phenotypes of cancers and their interacting gene and protein pathways to improve outcomes.
Development of a novel machine learning analysis technique
---For example, he used multiplicity measures of cancer gene networks to organize information in a clinically meaningful manner by integrating data from thousands of experiments testing for tumor somatic mutations. Using multiplicity measures of somatic mutations to organize cancers, he found that somatic mutations with high multiplicity scores correlated with known causal germline mutations.
---Extending techniques related to predicting glycan structure through gene expression data.
Developing a novel data integration approach using ontological representations to combine data from multiple experiments
---The approach has been applied to characterize dendritic cell maturation relevant to cancer immunotherapy.