Event Schedule

Monday, July 20 — Introduction, Protein Science & Installs

Morning: Welcome & Foundations

• Welcome and introductions with Prof. Jeff Gray and the teaching team; bootcamp goals, structure, norms, and cohort formation

• Introduction to Protein Design: the modern design pipeline and why the order of steps matters

• Protein Science Basics: amino acids, secondary and tertiary structure, what makes a protein "designed"

• Brief Intro to PyMol (~15 min): loading structures, navigating chains and residues, visualizing surfaces

• Note: no standalone Python Basics session — this cohort arrives Python-proficient

• Resources: [ Protein Design slides ] · [ Protein Science slides ] · [ PyMol slides ]

Afternoon: Installation & Cluster Day

• Install the toolkit on your home cluster: RFdiffusion, ProteinMPNN, ESMFold (AlphaFold and RoseTTAFold run via Colab notebooks)

• HPC / cluster orientation: logging in, environments, submitting jobs, command-line basics

• Experienced students pair with those still setting up — teaching it is the best way to learn it

• Resources: [ Install guide / Homework #1 ] · [ Cluster login sheet ] · [ Colab notebooks (AlphaFold, RoseTTAFold) ]

Evening: Intro to ML — Lecture & Tutorial

• Core ML concepts plus the transformer architecture that underpins most current tools

• Hands-on tutorial

• Resources: [ Intro to ML slides ] · [ Classification notebook ] · [ Neural network notebook ]

Study Hall: open study time with instructor support

Before you arrive (Homework #1): confirm cluster access and complete the install-status survey — [ link ]

Tuesday, July 21 — Structure Prediction & Backbone Generation

Morning: Structure Prediction — Concepts & Tools

• How structure prediction became tractable and how AlphaFold changed the field

• Tool survey: AlphaFold2/3, Boltz-1/2, Chai-1 — trade-offs and use cases

• Interpreting outputs: pLDDT, PAE, and common failure modes

• Resources: [ AF slides ] · [ AF notebook ] · [ AF alternatives slides ]

Afternoon: Backbone Generation with RFdiffusion

• Diffusion models conceptually; RFdiffusion key parameters and setting up a generation campaign

• CPU vs. GPU, HPC, and Bash scripting mini-lesson

• Resources: [ RFdiffusion slides ] · [ RFdiffusion notebook ] · [ CPU vs GPU slides ] · [ Bash scripting slides ]

Evening: RFdiffusion Tutorial

• Run backbone generation, inspect outputs, build intuition by comparing results; Boltz & Chai tutorial work

• Resources: [ RFdiffusion notebook ] · [ Data visualization slides/notebook ]

Study Hall: open study time with instructor support

Wednesday, July 22 — Sequence Design, Language Models & Rosetta Scoring

Morning: Sequence Design — ProteinMPNN & Protein Language Models

• Sequence-structure compatibility: what makes a sequence fold as intended

• ProteinMPNN: how it works, how to run it, what the output score means

• Protein Language Models: ESM2/ESM3 as encoders vs. generative models; zero-shot mutation scoring

• When to use structure-based vs. language-model approaches

• Resources: [ ProteinMPNN slides ] · [ ESM slides ] · [ ESM notebook ]

Afternoon: Rosetta Pose & Scoring — Bringing It All Together

• The Rosetta Pose object: what it is and how Rosetta uses it (folded in here alongside scoring)

• Using Rosetta to evaluate ML-generated designs; score terms that matter (interface energy, shape complementarity, buried unsatisfied H-bonds)

• Filtering and ranking a design set: from many candidates to a short experimental list

• Conda tutorial: environments, dependencies, reproducible workflows

• Resources: [ Rosetta scoring slides ] · [ Conda slides ] · [ Interface Analyzer docs ]

Evening: Scoring Tutorial

• Apply Rosetta scoring to the designs generated earlier in the week

• Discussion: connecting Rosetta scores to experimental risk — what does a "good" score actually tell us?

• Resources: [ Scoring notebook ] · [ Interface Analyzer notebook ] · [ Practice PDB ]

Study Hall: open study time with instructor support

Thursday, July 23 — Advanced Tools & Design Project

Morning: Tool Spotlights, Journal Club & Project Kickoff

• TA-led tool spotlights / extended lectures (e.g., ESMFold2, Boltz deep-dive, feature analysis) — leveraging the teaching team's specialties

• Journal-club-style discussion of 1–2 assigned papers (assigned as homework)

• Data Visualization: analyzing and communicating design results

• Project Introduction & Planning: design challenge framing, target assignment, group strategy

• QM & Enzyme Design is optional this year (not an NSF requirement for this cohort) — offered as a short talk or swapped for deeper tool coverage / project time

• Resources: [ Tool spotlight slides ] · [ Assigned papers ] · [ Data viz slides ] · [ Project intro slides ]

Afternoon: Project Work

• Groups run the full pipeline: structure prediction → backbone generation → sequence design → Rosetta scoring

• Several worked examples provided so groups can branch beyond a single template

• Instructor and TA support throughout; iterative refinement

• Resources: [ Project folder with notebooks ]

Evening: Extended Project Work

• Continue running and evaluating pipelines; optional: refold designs with Boltz/Chai and compare to AlphaFold

• Resources: [ Project folder ]

Study Hall: open study time with instructor support

Friday, July 24 — Presentations & Closing

Morning: Presentation Preparation

• Groups finalize outputs and prepare 10–12 minute presentations: design target & strategy, pipeline decisions, results & scores, limitations, proposed next steps

• Peer review: exchange notebooks or slides and give feedback

Afternoon: Group Presentations & Closing

• Group presentations with Q&A from instructors and peers

• Closing discussion: what did we learn, what surprised us, what would we do differently?

• Introduction to the Rosetta Commons: community resources and opportunities

• Next steps: recommended readings, tools to explore, and how to build on the week

• Resources: [ Papers & resources page ]

Evening: Free — well earned!