doi.bio/brian_hie

Brian Hie

Early Life and Education

Brian Hie is an Assistant Professor of Chemical Engineering and Data Science at Stanford University and an Innovation Investigator at Arc Institute. Hie was born and raised in the United States. During his college years, Hie was passionate about poetry, particularly the work of 17th-century poet John Donne. When it came time to choose a graduate school, Hie was torn between pursuing English at Harvard and computer science at MIT. Ultimately, he chose computer science.

Career

Hie is currently an Assistant Professor of Chemical Engineering and Data Science at Stanford University and an Innovation Investigator at Arc Institute. He supervises the Laboratory of Evolutionary Design, which conducts research at the intersection of biology and machine learning.

Previously, Hie was a Stanford Science Fellow in the Stanford University School of Medicine and a Visiting Researcher at Meta AI. He completed his Ph.D. at MIT CSAIL and was an undergraduate at Stanford University. He has also worked at Google X, Illumina, and Salesforce.

Research and Publications

Hie's research focuses on the intersection of biology and machine learning, specifically using machine learning techniques derived from natural language processing to understand protein evolution. He has published extensively in this field, with notable publications including:

• "Efficient evolution of human antibodies from general protein language models" (2023)
• "Machine Learning for Protein Engineering" (2023)
• "Evolutionary-scale prediction of atomic-level protein structure with a language model" (2023)
• "Learning the language of viral evolution and escape" (2021)
• "Leveraging Uncertainty in Machine Learning Accelerates Biological Discovery and Design" (2020)

- "Efficient integration of heterogeneous single-cell transcriptomes using Scanorama" (2019)

Brian Hie

Brian Hie is an Assistant Professor of Chemical Engineering and Data Science at Stanford University and an Innovation Investigator at the Arc Institute. He supervises the Laboratory of Evolutionary Design, which conducts research at the intersection of biology and machine learning.

Education

Hie completed his undergraduate degree at Stanford University and went on to obtain a Ph.D. from the Massachusetts Institute of Technology (MIT) in Electrical Engineering and Computer Science.

Career

Before joining Stanford University as a faculty member, Hie worked at Google X, Illumina, and Salesforce. He also held positions as a Stanford Science Fellow at the Stanford University School of Medicine and a Visiting Researcher at Meta AI.

Research

Hie's research focuses on the application of machine learning techniques, particularly natural language processing, to the field of biology. He uses "protein language models" trained on large repositories of protein sequences to understand and predict protein evolution. His work has demonstrated that algorithms can predict protein evolution over extended periods, and he is now exploring the use of these models to design new proteins artificially.

Notable Works

• Efficient evolution of human antibodies from general protein language models – This work demonstrates how general protein language models can efficiently evolve human antibodies by suggesting evolutionarily plausible mutations, even without information about the target antigen or protein structure.
• Evolutionary velocity with protein language models predicts evolutionary dynamics of diverse proteins – In this paper, Hie and his co-authors introduce the concept of "evo-velocity," a dynamic vector field of protein evolution. This approach allows for predicting the evolutionary dynamics of proteins over vastly different timescales and provides insights into viral-host immune escape and the evolution of specific protein families.
• Learning the language of viral evolution and escape – Here, Hie and his colleagues use machine learning algorithms originally developed for human natural language to model viral escape, i.e., viral mutations that evade the immune system and impede vaccine development.
• Leveraging Uncertainty in Machine Learning Accelerates Biological Discovery and Design – This paper discusses the benefits of quantifying prediction uncertainty in machine learning algorithms to handle novel phenomena and facilitate biological discovery.
• Geometric Sketching Compactly Summarizes the Single-Cell Transcriptomic Landscape – Hie and his co-authors propose a method to enhance and accelerate single-cell data analysis by summarizing transcriptomic heterogeneity within a dataset, enabling more comprehensive visualization and sensitive detection of rare cell types.
• Efficient integration of heterogeneous single-cell transcriptomes using Scanorama – Scanorama is an algorithm presented by Hie and colleagues to integrate single-cell RNA sequencing (scRNA-seq) data from multiple experiments, laboratories, and technologies. It enables the merging of shared cell types and the removal of batch effects, facilitating the discovery of biological insights.
• Realizing private and practical pharmacological collaboration – In this work, Hie et al. introduce a computational protocol that ensures data confidentiality while allowing for secure and collaborative training of predictive models in pharmacological research, thus enhancing the potential for life-saving breakthroughs.
• Pooled ChIP-Seq Links Variation in Transcription Factor Binding to Complex Disease Risk – The paper introduces a pooling-based approach to mapping quantitative trait loci (QTLs) for molecular-level traits, specifically focusing on transcription factor (TF) binding sites and a histone modification. This method reduces costs significantly and provides insights into genetic variants affecting TF binding and chromosomal architecture.
• Schema: metric learning enables interpretable synthesis of heterogeneous single-cell modalities – Schema is a framework that uses metric learning to synthesize information across heterogeneous modalities, such as age, disease status, gene expression, and chromatin accessibility, to gain a comprehensive understanding of biological processes.
• Predicting the mutational drivers of future SARS-CoV-2 variants of concern – In this study, Hie and his co-authors aim to predict Spike amino acid changes that could contribute to future variants of concern in SARS-CoV-2. They test the importance of features such as epidemiology, evolution, immunology, and neural network-based protein structure predictions.

Youtube Videos

Youtube Title: AIRR-C Seminar Series, February 22nd, 2024 - Brian Hie, Stanford University, US

Youtube Link: link

Youtube Channel Name: AIRR Community

Youtube Channel Link: https://www.youtube.com/@AIRRCommunity

AIRR-C Seminar Series, February 22nd, 2024 - Brian Hie, Stanford University, US

Youtube Title: Scientist Stories: Brian Hie, Evolution of Human Antibodies From General Protein Language Models

Youtube Link: link

Youtube Channel Name: Axial

Youtube Channel Link: https://www.youtube.com/@axialxyz

Scientist Stories: Brian Hie, Evolution of Human Antibodies From General Protein Language Models

Youtube Title: Evo: DNA foundation modeling from molecular to genome scale | Brian Hie

Youtube Link: link

Youtube Channel Name: VantAI

Youtube Channel Link: https://www.youtube.com/@Vant_AI

Evo: DNA foundation modeling from molecular to genome scale | Brian Hie

Youtube Title: MIA: Brian Hie, Learning to read and write protein evolution

Youtube Link: link

Youtube Channel Name: Broad Institute

Youtube Channel Link: https://www.youtube.com/@broadinstitute

MIA: Brian Hie, Learning to read and write protein evolution

Youtube Title: Matthew Moments | Episode 3 | Brian Hie

Youtube Link: link

Youtube Channel Name: FirstSF

Youtube Channel Link: https://www.youtube.com/@FirstSF

Matthew Moments | Episode 3 | Brian Hie

Youtube Title: Scientist Stories: Brian Hie, Learning to Read & Write Protein Evolution

Youtube Link: link

Youtube Channel Name: Axial

Youtube Channel Link: https://www.youtube.com/@axialxyz

Scientist Stories: Brian Hie, Learning to Read & Write Protein Evolution

Youtube Title: Efficient Evolution of Human Antibodies From General Protein Language Models and Sequence Info Alone

Youtube Link: link

Youtube Channel Name: ML for protein engineering seminar series

Youtube Channel Link: https://www.youtube.com/@mlforproteinengineeringsem6420

Efficient Evolution of Human Antibodies From General Protein Language Models and Sequence Info Alone

Youtube Title: VBS is too short - Brian Hie (Parody of "Baby")

Youtube Link: link

Youtube Channel Name: Darren Tung

Youtube Channel Link: https://www.youtube.com/@dtung118

VBS is too short - Brian Hie (Parody of "Baby")

Youtube Title: Biological modeling from molecular to genome scale | AI and the Molecular World | Brian Hie

Youtube Link: link

Youtube Channel Name: Applied Machine Learning Days

Youtube Channel Link: https://www.youtube.com/@AppliedMachineLearningDays

Biological modeling from molecular to genome scale | AI and the Molecular World | Brian Hie

Youtube Title: kgml2021: Translational Biology, Brian Hie, Stanford University

Youtube Link: link

Youtube Channel Name: KGML Workshop

Youtube Channel Link: https://www.youtube.com/@kgmlworkshop6368

kgml2021: Translational Biology, Brian Hie, Stanford University

Youtube Title: HIE Research Update with Dr. Brian Kalish

Youtube Link: link

Youtube Channel Name: Hope for HIE

Youtube Channel Link: https://www.youtube.com/@HopeforHIE

HIE Research Update with Dr. Brian Kalish

Youtube Title: Scientist Stories: Eric Nguyen, Using AI to Design from the molecular to genome scale

Youtube Link: link

Youtube Channel Name: Axial

Youtube Channel Link: https://www.youtube.com/@axialxyz

Scientist Stories: Eric Nguyen, Using AI to Design from the molecular to genome scale

Youtube Title: HIE Awareness Month Member stories

Youtube Link: link

Youtube Channel Name: Hope for HIE

Youtube Channel Link: https://www.youtube.com/@HopeforHIE

HIE Awareness Month Member stories

Youtube Title: [Summary] The Gamechangers Webinar 6 | Disruptive Methodologies: AI, Machine Learning, and AMR

Youtube Link: link

Youtube Channel Name: RADAAR Team

Youtube Channel Link: https://www.youtube.com/@radaarteam513

![Summary] The Gamechangers Webinar 6 | Disruptive Methodologies: AI, Machine Learning, and AMR

Youtube Title: Studies underway to improve brain health in oxygen-deprived newborns

Youtube Link: link

Youtube Channel Name: UT Southwestern Medical Center

Youtube Channel Link: https://www.youtube.com/@UTSWMed

Studies underway to improve brain health in oxygen-deprived newborns

Youtube Title: EVO: DNA Foundation Models - Eric Nguyen | Stanford MLSys #96

Youtube Link: link

Youtube Channel Name: Stanford MLSys Seminars

Youtube Channel Link: https://www.youtube.com/@StanfordMLSysSeminars

EVO: DNA Foundation Models - Eric Nguyen | Stanford MLSys #96

Youtube Title: The Gamechangers | RADAAR AMR Policy Webinar-6

Youtube Link: link

Youtube Channel Name: RADAAR Team

Youtube Channel Link: https://www.youtube.com/@radaarteam513

The Gamechangers | RADAAR AMR Policy Webinar-6

Youtube Title: If You Could Hie to Kolob

Youtube Link: link

Youtube Channel Name: Brian Daw

Youtube Channel Link: https://www.youtube.com/channel/UClm6XGOmS9X2vmfXwoZ-LNw

If You Could Hie to Kolob

Youtube Title: The Increasing Value of HIE in Colorado

Youtube Link: link

Youtube Channel Name: Contexture

Youtube Channel Link: https://www.youtube.com/@contexture2065

The Increasing Value of HIE in Colorado