January 2, 2019
Amanda Minnich, Ph.D., ATOM Data Scientist, is spearheading the ATOM Data Driven Modeling (DDM) Capability Team and has led the development of our automated DDM pipeline. Amanda has been an active contributor to outreach efforts such as Women in HPC events and Lawrence Livermore National Laboratory’s video production of, “What is Machine Learning?”
Amanda joined Lawrence Livermore National Lab in July 2017 as a Research Scientist, and, shortly after, joined ATOM. As part of the ATOM Team, she applies machine learning (ML) techniques to biological data for drug discovery purposes.
This past November, Amanda presented an update on ATOM’s computational framework at the Fourth Computational Approaches for Cancer Workshop at SuperComputing 2018 in Dallas, Texas. This framework is currently capable of building machine learning models that generate all key safety and pharmacokinetics (PK) parameters needed as input for ATOM’s Quantitative System Pharmacology and Toxicology models. Our end-to-end pipeline first ingests raw datasets, curates them, and stores the result in our data lake. Next, it extracts features from these data and trains and saves the model to our model zoo. Our pipeline generates a variety of molecular features and both shallow and deep ML models. The HPC-specific module we have developed conducts efficient parallelized search of the model hyperparameter space and reports the best-performing hyperparameters for each of these feature/model combinations.
Amanda’s team’s contributions have laid the foundation for application of the pipeline on pilot projects in our second year, as we look to integrate drug efficacy modeling in addition to safety and PK. Our models are currently being integrated into an active learning pipeline to aid in de novo compound generation, as well as being sent back to consortium members to incorporate into their drug discovery efforts. Our models and code will also be released to the public after our consortium member benefit period.