Brian Koepnick
A5063058216- Protein Structure and Dynamics
- RNA and protein synthesis mechanisms
- Enzyme Structure and Function
- Teaching and Learning Programming
- Machine Learning in Bioinformatics
- Mass Spectrometry Techniques and Applications
- Genetics, Bioinformatics, and Biomedical Research
- Biomedical and Engineering Education
- Wikis in Education and Collaboration
- Integrated Circuits and Semiconductor Failure Analysis
- Pharmacological Effects of Medicinal Plants
- Educational Games and Gamification
- Biochemical and Structural Characterization
- RNA Research and Splicing
- Nuclear Structure and Function
- Bacteriophages and microbial interactions
- Fuel Cells and Related Materials
- Semiconductor materials and devices
- Advanced Electron Microscopy Techniques and Applications
- Ubiquitin and proteasome pathways
- ATP Synthase and ATPases Research
- Electron and X-Ray Spectroscopy Techniques
- Photosynthetic Processes and Mechanisms
- Medicinal Plant Pharmacodynamics Research
- Transgenic Plants and Applications
University of Washington
2016-2024
Seattle University
2019
Although deep learning has revolutionized protein structure prediction, almost all experimentally characterized de novo designs have been generated using physically based approaches such as Rosetta. Here, we describe a learning-based sequence design method, ProteinMPNN, that outstanding performance in both silico and experimental tests. On native backbones, ProteinMPNN recovery of 52.4% compared with 32.9% for The amino acid at different positions can be coupled between single or multiple...
Significance Almost all proteins fold to their lowest free energy state, which is determined by amino acid sequence. Computational protein design has primarily focused on finding sequences that have very low in the target designed structure. However, what most relevant during folding not absolute of folded state but difference between and lowest-lying alternative states. We describe a deep learning approach captures aspects landscape, particular presence structures minima, show it can...
Abstract While deep learning has revolutionized protein structure prediction, almost all experimentally characterized de novo designs have been generated using physically based approaches such as Rosetta. Here we describe a sequence design method, ProteinMPNN, with outstanding performance in both silico and experimental tests. The amino acid at different positions can be coupled between single or multiple chains, enabling application to wide range of current challenges. On native backbones,...
We show here that computer game players can build high-quality crystal structures. Introduction of a new feature into the Foldit allows to and real-space refine structures electron density maps. To assess usefulness this feature, we held crystallographic model-building competition between trained crystallographers, undergraduate students, automatic algorithms. After removal disordered residues, team achieved most accurate structure. Analysing target protein competition, YPL067C, uncovered...
Abstract The protein design problem is to identify an amino acid sequence which folds a desired structure. Given Anfinsen’s thermodynamic hypothesis of folding, this can be recast as finding for the lowest energy conformation that As calculation involves not only all possible sequences but also structures, most current approaches focus instead on more tractable structure, often checking by structure prediction in second step indeed designed sequence, and discarding many cases large fraction...
Despite progress in the design of protein binding proteins, shape matching binder to target has not yet reached that highly evolved native protein-protein complexes, and previous efforts have failed for hard targets such as TNF receptor (TNFR1) relatively flat polar surfaces. We reasoned free diffusion starting from random noise could enable generation extensive shape-matching binders challenging targets, tested this approach on TNFR1 related super family members. The diffused nanomolar...
Abstract The computer game Foldit is currently widely used as a biology and biochemistry teaching aid. Herein, we introduce new feature of called “custom contests” that allows educators to create puzzles fit their curriculum. effectiveness the custom contests demonstrated by use five distinct in an upper‐level class. contest can be implemented classes ranging from middle school graduate enable best complement current © 2019 International Union Biochemistry Molecular Biology, 47(2): 133–139, 2019.
Despite progress in designing protein-binding proteins, the shape matching of designs to targets is lower than many native protein complexes, and design efforts have failed for tumor necrosis factor receptor 1 (TNFR1) other with relatively flat polar surfaces. We hypothesized that free diffusion from random noise could generate shape-matched binders challenging tested this approach on TNFR1. obtain low picomolar affinity whose specificity can be completely switched family members using...
Abstract Every two years groups worldwide participate in the Critical Assessment of Protein Structure Prediction (CASP) experiment to blindly test strengths and weaknesses their computational methods. CASP has significantly advanced field but many hurdles still remain, which may require new ideas collaborations. In 2012 a web-based effort called WeFold, was initiated promote collaboration within community attract researchers from other fields contribute CASP. Members WeFold coopetition...
With the rapid improvement of cryo-electron microscopy (cryo-EM) resolution, new computational tools are needed to assist and improve upon atomic model building refinement options. This communication demonstrates that microscopists can now collaborate with players computer game Foldit generate high-quality de novo structural models. development could greatly speed generation excellent cryo-EM structures when used in addition current methods.
Undergraduate research experiences can improve student success in graduate education and STEM careers. During the COVID-19 pandemic, undergraduate researchers at our institution many others lost their work–study positions due to interruption of in-person activities. This imposed a financial burden on students eliminated an important learning opportunity. To address these challenges, we created paid, fully remote, cohort-based curriculum computational protein design. Our used existing design...
Abstract Undergraduate research experiences can improve student success in graduate education and STEM careers. During the COVID-19 pandemic, undergraduate researchers at our institution many others lost their work-study positions due to interruption of in-person activities. This imposed a financial burden on students eliminated an important learning opportunity. To address these challenges, we created paid, fully-remote, cohort-based curriculum computational protein design. Our used...