Skip to main content

Interpretable Machine Learning for Multivariate Omics Data Analysis


EMSL Project ID
60480

Abstract

Often, the ultimate motivation for the generation of omics data for an experiment is a better understanding of an observed phenotype of a biological system and the effect of perturbations on the phenotype. In analyzing omics data, traditional univariate statistical models (i.e. each biomolecule is analyzed separately) are typically run and then summarized for inference by methods such as enrichment analysis. However, univariate statistical methods alone are not sufficient for identifying complex biochemical processes or revealing multigene phenotypes. Machine learning (ML) provides a complementary suite of methods for evaluating complex multivariate relationships. We propose to develop a web application for EMSL users to perform reproducible and robust ML methods on omics data with guided selection of models and parameters and computational and visualization tools to aid in the interpretation of results within the biological context.

Project Details

Start Date
2022-10-01
End Date
N/A
Status
Active

Team

Principal Investigator

Lisa Bramer
Institution
Pacific Northwest National Laboratory

Team Members

David Degnan
Institution
Pacific Northwest National Laboratory

Daniel Claborne
Institution
Pacific Northwest National Laboratory

Sneha Couvillion
Institution
Pacific Northwest National Laboratory

Kelly Stratton
Institution
Pacific Northwest National Laboratory

Kristin Burnum-Johnson
Institution
Environmental Molecular Sciences Laboratory