Noah Flynn

Applied Scientist II at Amazon AWS AI Labs | PhD Deep Learning & Cheminformatics | Adjunct Faculty at UC Berkeley

XY [AT] wustl.edu (X=noah, Y=flynn)

News

I wrote a book with Manning Publications! I'm happy to announce that "Machine Learning for Drug Discovery" is now available via Manning's Early Access Program (MEAP)! Check it out here and use code "au35fly" for a 35% discount! Feel free to browse (or contribute) to the official code repository, though keep in mind that significant development is still underway! Happy learning!

The search for novel, improved therapeutics is an important problem and impacts all of our lives. "Machine Learning for Drug Discovery" is an informative deep dive that aligns machine learning and deep learning theory with contemporary applications in drug discovery. While a cliché, it is nevertheless true: this is the book I wish I had available to me when I started my graduate studies and early career in industry. Each chapter is structured around real-world scenarios and iteratively builds up a solution with code provided every step of the way. By the end, you'll have the agency to explore, apply, replicate, and improve upon innovations and publications at the intersection of machine learning and drug discovery. Both seasoned researchers and newcomers welcome -- no prior background in chemistry or machine learning is necessary!

I've recently released two more chapters! In chapter 6, "Cast Study: Small Molecule Binding to an RNA Target," we introduce RNA's role in therapeutic development by replicating recent computational studies from Cai et al. [1] and Yazdani et al [2], contextualizing the authors' decisions step-by-step and producing an exemplary quantitative structure-activity relationship (QSAR) pipeline. In contrast, chapter 7, "Unsupervised Learning: Repurposing Drugs, Curating Compounds, & Screening Fragments," is more like a buffet of unsupervised learning applications in drug repurposing (dimensionality reduction), compound library design (clustering), and fragment-based drug discovery (density estimation), with the latter based on work from McCorkindale et al. [3]

Bio

Hello! I'm Noah, a research scientist at Amazon within AWS AI Labs. Previously, I was a member of Amazon's Artificial General Intelligence (AGI) organization. I am also adjunct faculty within UC Berkeley's Master of Molecular Science and Software Engineering program, where I am the course designer and instructor for "Software Engineering Fundamentals for Molecular Sciences." My focus is on applications of large language models in dialogue systems, multilingual natural language understanding, and graph machine learning for natural language processing.

I also maintain expertise and interest in advancing the field of machine learning and drug discovery. Prior to Amazon, I earned a PhD in Computational Biology at Washington University School of Medicine in St. Louis within the Department of Pathology & Immunology (Thesis). My research focused on applications of deep learning to improving drug toxicity screening and safety assessment. For example, detecting signals from electronic health records that relate to drug-drug interactions and modeling drug metabolism networks to predict vectors of drug toxicity, formation of reactive chemical species, and likelihood of adverse drug reactions. As part of my research, I served as a core maintainer of XenoSite and I launched the XenoNet Web Server. XenoSite is a prediction web service for small molecule biochemistry that hosts an online collection of metabolism and reactivity mathematical models that systematically summarize the data from thousands of papers into a condensed and useable computational tool that predicts adduction and potential toxicity of molecules.

If you are looking for a collaborator or simply want to connect with someone who shares your passion for using technology to make a positive impact in the world, please don't hesitate to get in touch.

I am in the process of updating my website and have a few upcoming projects to announce heading into 2023 Q3, so feel free to keep checking in!

Publications

Most recent publications on Google Scholar.
‡ indicates equal contribution.

SEACrowd: A Multilingual Multimodal Data Hub and Benchmark Suite for Southeast Asian Languages

Holy Lovenia, et al. (including Noah R. Flynn)

EMNLP (2024)

Paper

Message Passing Neural Networks Improve Prediction of Metabolite Authenticity

Noah R Flynn, S Joshua Swamidass

Journal of Chemical Information and Modeling (2023)

Paper Supporting Information

Discovery of Novel Reductive Elimination Pathway for 10-Hydroxywarfarin

Dakota L Pouncey, et al. (including Noah R Flynn)

Frontiers in Pharmacology (2022)

Paper

Cal-net: Jointly learning classification and calibration on imbalanced binary classification tasks

Arghya Datta, Noah R Flynn, S Joshua Swamidass

IJCNN (2021)

Paper

Bioactivation of Isoxazole-Containing Bromodomain and Extra-Terminal Domain (BET) Inhibitors

Noah Flynn, Michael D. Ward, Mary A. Schleiff, et al.

Metabolites (2021)

Paper Supporting Information

Machine Learning on Liver-Injuring Drug Interactions with NSAIDs from Hospitalization Data

Arghya Datta‡, Noah Flynn‡, Dustyn A Barnette, et al.

PLOS Computational Biology (2021)

Paper

Modeling the Bioactivation and Subsequent Reactivity of Drugs

Tyler B Hughes‡, Noah Flynn‡, Na Le Dang, S Joshua Swamidass

Chemical Research in Toxicology (2021)

Paper

Significance of Multiple Bioactivation Pathways for Meclofenamate as Revealed through Modeling and Reaction Kinetics

Mary A Schleiff, Noah R Flynn, Sasin Payakachat, et al.

Drug Metabolism and Disposition (2021)

Paper Supporting Information

Meloxicam methyl group determines enzyme specificity for thiazole bioactivation compared to sudoxicam

Dustyn A Barnette, et al. (including Noah Flynn)

Toxicology Letters (2020)

Paper

The Metabolic Forest: Predicting the Diverse Structures of Drug Metabolites

Tyler B Hughes, Na Le Dang, Ayush Kumar, Noah R Flynn, S Joshua Swamidass

Journal of Chemical Information and Modeling (2020)

Paper

XenoNet: Inference and Likelihood of Intermediate Metabolite Formation

Noah R Flynn, Na Le Dang, Michael D Ward, S Joshua Swamidass

J. of Chem. Inf. Model. (2020)

Paper Supporting Information

Dual mechanisms suppress meloxicam bioactivation relative to sudoxicam

Dustyn A Barnette, et al. (including Noah Flynn)

Toxicology (2020)

Paper

Comprehensive kinetic and modeling analyses revealed CYP2C9 and 3A4 determine terbinafine metabolic clearance and bioactivation

Dustyn A Barnette, Mary A Davis, Noah Flynn, et al.

Biochem. Pharmacol. (2019)

Paper

CYP2C19 and 3A4 Dominate Metabolic Clearance and Bioactivation of Terbinafine Based on Computational and Experimental Approaches

Mary A Davis, Dustyn A Barnette, Noah R Flynn, et al.

Chemical Research in Toxicology (2019)

Paper Supporting Information

Message Passing Neural Networks Improve Prediction of Metabolite Authenticity

Noah R Flynn, S Joshua Swamidass

Journal of Chemical Information and Modeling (2023)

Paper Supporting Information

Cal-net: Jointly learning classification and calibration on imbalanced binary classification tasks

Arghya Datta, Noah R Flynn, S Joshua Swamidass

IJCNN (2021)

Paper

Bioactivation of Isoxazole-Containing Bromodomain and Extra-Terminal Domain (BET) Inhibitors

Noah Flynn, Michael D. Ward, Mary A. Schleiff, et al.

Metabolites (2021)

Paper Supporting Information

Machine Learning on Liver-Injuring Drug Interactions with NSAIDs from Hospitalization Data

Arghya Datta‡, Noah Flynn‡, Dustyn A Barnette, et al.

PLOS Computational Biology (2021)

Paper

Modeling the Bioactivation and Subsequent Reactivity of Drugs

Tyler B Hughes‡, Noah Flynn‡, Na Le Dang, S Joshua Swamidass

Chemical Research in Toxicology (2021)

Paper

XenoNet: Inference and Likelihood of Intermediate Metabolite Formation

Noah R Flynn, Na Le Dang, Michael D Ward, S Joshua Swamidass

J. of Chem. Inf. Model. (2020)

Paper Supporting Information

Vitæ

Full Resume in PDF.

Website Design

This website was built with Jekyll based on a template by Martin Saveski.
Noah Flynn
XY [AT] wustl.edu (X=noah, Y=flynn)