|
Abstract:
|
Arivale is a consumer-facing scientific wellness company that collects blood, saliva, stool, and survey data at regular intervals from individuals who enroll in the Arivale program. The biosamples are analyzed to collect a wide range of multi-omic data from each individual, including genomics, metabolomics, proteomics, gut microbiome composition, and clinical laboratory tests. Arivale also collects quantified-self data via wearable fitness trackers (e.g. Fitbit) and a battery of questionnaires on topics including health history, diet, and lifestyle. Using these personal, dense, dynamic data clouds collected from over 3500 individuals who consented for research, we perform network analysis via existing undirected graphical modeling methods in order to understand conditionally dependent relationships within Arivale's vast collection of data. With the discovered networks, we vet known associations with physiology and disease, propose novel connections, and discuss the challenges and limitations in applying existing statistical methodology to Arivale's unique collection of data.
|
