Statistical Learning for Marine Ecosystems

Investigator: Jacob Bien, USC

Oceanographers have assembled a wide range of techniques for measuring marine ecosystems. Having many different data types, or “views” of these ecosystems, promises to provide a more nuanced picture of the inner workings of the complex underlying systems. However, realizing this promise requires statistical methods that can reliably integrate these myriad data sets together in a principled, reproducible and computationally feasible way that will highlight important relationships.  We focus on developing new statistical methodology in collaboration with oceanographers in CBIOMES.

Our work revolves primarily around three main data types: (1) flow cytometry data, (2) taxonomically-structured microbial amplicon data, and (3) environmental data (especially those collected via satellite).  Being able to explicitly tie these diverse data types together provides the opportunity for new insights.  Our work will involve the development of new statistical methodology, data-focused collaboration, and implementation and distribution of open-source software.

CBIOMES Collaborators in Bien’s Group

News from the Bien Group

Past Annual Reporting

2020 | 2019 | 2018 [login required]