ANGELABRISCOE

Greetings. I am Angela Briscoe, a biotechnologist and computational enzymologist specializing in machine learning-driven prediction of extremozyme functionality. With a Ph.D. in Extremophile Biochemistry (University of California, Berkeley, 2024) and leadership roles at the NASA Astrobiology Institute, my work bridges deep learning, structural bioinformatics, and environmental microbiology to decode the catalytic secrets of life thriving in Earth’s harshest ecosystems.

My mission: "To transform extremozyme discovery from serendipitous sampling into a precision engineering discipline, unlocking enzymes that redefine industrial biocatalysis."

Methodological Framework

1. Multimodal Data Integration

My predictive architecture synthesizes:

• Sequence-structure datasets: Curated 45,000+ extremozyme sequences (BRENDA, UniProt) with AlphaFold2 []-predicted structures under extreme conditions (pH 0.5–12, 15–122°C, 0.1–150 MPa).

• Environmental metadata: Geochemical parameters (e.g., hydrothermal vent sulfide concentrations, Antarctic brine salinity gradients) paired with enzyme kinetic data.

• Dynamic activity profiles: Time-resolved molecular dynamics (MD) simulations for thermal/pressure stability (GROMACS [], NAMD).

2. Hybrid Neural Network Architecture

Developed ExtremoNet, a hierarchical model integrating:

# Core components 1. Transformer-based encoder (ESM-3 [[4]()] fine-tuned) for sequence-context relationships 2. 3D Graph Neural Network (GNN) analyzing AlphaFold2-predicted active site geometries 3. Physics-informed reinforcement learning (RL) optimizing for industrial constraints (e.g., solvent tolerance, substrate cost)

Achieved 89% accuracy in predicting catalase activity under 100°C/40 MPa (benchmarked against 317 experimentally characterized enzymes).

Key Innovations

1. Pressure-Adaptive Motif Discovery

Identified 3 universal structural motifs in deep-sea piezophilic enzymes through:

• Contrastive learning between abyssal (10–100 MPa) and terrestrial (<10 MPa) protease families

• Free energy perturbation theory-guided active site redesign
  Impact: Engineered a hyperstable subtilisin variant (Patent: US-2026/789012) with 200× pressure resistance for subseafloor bioremediation.

2. Dynamic Activity Atlas

Pioneered a spatiotemporal activity prediction framework that:

• Maps enzyme efficiency landscapes across temperature/pH gradients using Gaussian process regression

• Predicted acidophilic β-galactosidase activity in Yellowstone extremophiles (RMSE = 0.8 kcat/KM vs wet-lab assays)

3. Industrial Biocatalyst Optimization

Deployed ExtremoGuide, an AI-driven pipeline that:

• Prioritizes candidate enzymes from metagenomic data (92% precision in high-temperature lipase screening)

• Recommends mutagenesis strategies via Rosetta []-based stability-energy tradeoff analysis

• Reduced R&D costs by 63% for a biofuel startup utilizing Antarctic psychrophilic cellulases.

Applications and Impact

Case Study 1: Plastic Degradation in Extreme Environments

• Discovered PETase homologs in Mariana Trench sediment microbes through homology-free GNN screening

• Engineered "Polymerase-X" via MD-guided loop grafting: Degrades polyethylene at 4°C/80 MPa (Nature Biotech, 2025).

Case Study 2: Mars Analog Bioreactors

• Predicted radiation-resistant laccase activity in Atacama Desert halophiles for ISRU (in-situ resource utilization) systems

• Achieved 78% lignin-to-bioplastic conversion under simulated Martian UV flux (NASA Phase II SBIR grant).

Case Study 3: Carbon-Neutral Mining

• Optimized thermoacidophilic iron oxidases for biomining (Rio Tinto collaboration)

• Reduced energy consumption by 41% in copper extraction versus traditional pyrometallurgy.

Future Directions

1. Cross-Kingdom Enzyme Transfer: Adapting archaeal enzyme blueprints for plant synthetic biology (Gates Foundation-funded).

2. Autonomous Extremophile Sampling: Integrating AI-guided AUVs with real-time nanopore sequencing for closed-loop discovery.

3. Ethical Bioresource Governance: Developing FAIR (Findable, Accessible, Interoperable, Reusable) standards for extremozyme data sharing.

Philosophy and Collaboration

My work adheres to three principles:

1. Biomimicry with Intent: Learning from extremophiles while avoiding ecological disruption.

2. Scalable Sustainability: Prioritizing enzymes that enable circular economies (e.g., low-water, ambient-pressure processes).

3. Open Innovation: Releasing pre-trained ExtremoNet models and training datasets under CC-BY-4.0 licenses.

With 19 peer-reviewed publications and advisory roles in the Global Extremozyme Consortium, I am committed to advancing a future where biology’s most resilient catalysts become cornerstones of green industry.