Foundational virtual cells that are fully mechanistic
The finale. If a cell is binding and catalysis, and every tool in this course applies to that object, what stops us from writing the whole cell down and running it?
The course closes on the question it has been assembling all term. Whole-cell models exist and are hard: they are stitched from incompatible submodels, their parameters are unidentifiable, and they do not compose. The claim this lecture puts on the table is that a mechanistic foundational virtual cell becomes reachable if the whole cell is written in one language, and that binding-catalysis with reaction-order geometry is a candidate for that language. It is an open research programme, which makes it the right place to end: the students have every tool needed to argue about it.
This lecture is new in 2026 or substantially re-scoped, so parts of it are not yet carried over from the 2025 notes. The teaching team should treat this list as the commissioning brief, and the lecturer as the backlog.
- What a whole-cell model is today: the Karr 2012 E. coli/M. genitalium lineage, what it achieved and what it cost.
- Why current whole-cell models do not compose: submodel incompatibility, parameter unidentifiability, no shared reduction.
- The mechanistic foundational virtual cell proposal: one language for the whole cell, binding-catalysis as the candidate, reaction order as the shared structure. Provenance: the lecturer's foundational-virtual-cell project.
- Contrast with learned virtual cells: what a data-driven cell model can and cannot be asked.
- Close the arc: physics, system, industry, and what each school contributed to the object on the board.
The synthesis
Three schools, one proposal, and the arc of the course.
1Design a biomachine
The panel: given a function, propose a design using physics-costed, system-wired, industry-buildable components.
The integration panel is a design brief worked at the whiteboard. A function is proposed (a sensor that responds to a ligand with a burst; a switch with a memory; a consortium that fixes nitrogen and reports); the room decomposes it into: the physics constraints (numbers: k_on, copy numbers, energies, lectures 1, 2 and 7), the system structure (network, motifs, stability at the chosen timescales, noise budget, lectures 3 to 9), and the industry implementation (what would need designing and engineering, lectures 13 to 15). The deliverable is the proposal's PCAPS chain: problem, challenge, approach, proof, significance — one paragraph each.
2The course in one arc
OoM → reactions → dynamics → control → noise → energy → computation → industry: one machine, three lenses.
The arc, stated once: physics gave the scales and the costs
(lectures 1, 2, 7); system gave the structure — how a chemical network
shapes a state space, adapts, computes, decides (lectures 3 to 12); industry
gave the use — design proteins, engineer fluxes, grow machines (lectures 13 to 15). The final
research essay takes any topic you care about beyond the lectures, with the
research-essay pipeline behind it and a full PCAPS chain; the week-8
extension mini-essay is a good place to draft one.
One topic beyond the lectures; thesis you could disagree with; literature with verified full texts; numbers from code; 'what I would argue against'. Due two weeks after the last lecture; the teaching format used all semester (core → exposition → extension) is the reference shape.
†References & note sources
Where this page's claims and numbers live.
- F. Xiao, CCBS 2025 Lecture 9 — Introduction to Protein Design (slides: Rosetta, de novo binder pipeline, SARS-CoV-2/LCB1, RFdiffusion, ProteinMPNN, AlphaFold2/ColabFold/MMseqs2, hands-on recipes). PDF This lecture traces here, including the two run_inference recipes.
- F. Xiao, CCBS 2025 Lecture 11 — Growth machine (growth law, chemostat/Monod, consumer–resource, proteome partition λ = εφ_R, flux balance, upshift/downshift, Cobb–Douglas/Solow economics analogy, V̇ = λV frontier). PDF This lecture traces here.
- L. Cao et al., Science 2020 — SARS-CoV-2 miniprotein inhibitors (LCB1). The case study for this lecture.
- J. Watson et al., Nature 2023 — RFdiffusion; J. Ingraham et al., Science 2023 — ProteinMPNN; J. Jumper et al., Nature 2021 — AlphaFold2; M. Steinegger & J. Söding, Nat. Biotechnol. 2017 — MMseqs2. Toolkit references (cited on the lecture slides).
- M. Scott, C. Gunderson, E. Mateescu, Z. Zhang, T. Hwa, Science 2011 — the proteome-partition growth law.The backbone result for this lecture.
- R. Phillips & J. Kondev, Cell biology by the numbers (book.bionumbers.org) — metabolic yields and energy scales.The numbers for this lecture.