Define systems biology and holistic modeling

WHY does this field exist?

For a century biology worked by reductionism: smash the organism into pieces (genes, proteins, metabolites), study each piece alone, publish. This gave us the parts catalogue (genome, proteome...).

But knowing every brick does not tell you the shape of the house. A heartbeat, a circadian rhythm, an oscillating gene circuit — these are dynamic behaviours produced by the wiring, not by any one molecule.

WHAT is systems biology?

HOW do we build a holistic model? (derive from scratch)

Let's derive the simplest holistic model: a network of interacting concentrations.

Step 1 — State variables. Let the system have $n$ components with concentrations $\mathbf{x}(t) = (x_1, x_2, \dots, x_n).$ Why this step? You cannot model change without first naming what changes. Concentration is the natural variable for molecules.

Step 2 — Rate of change = production − consumption. For each component, mass balance says $\frac{dx_i}{dt} = (\text{rate in}) - (\text{rate out}).$ Why this step? This is just conservation of matter — the only thing that changes a pool is stuff entering or leaving.

Step 3 — Interactions make the rates depend on other components. In a network, $x_i$'s production may depend on $x_j$. So: $\boxed{\dfrac{dx_i}{dt} = f_i(x_1, x_2, \dots, x_n)}$ Why this step? The coupling of $f_i$ to other variables is exactly what "network" means. If every $f_i$ depended only on $x_i$, there would be no system — just $n$ separate problems.

Step 4 — Concrete interaction terms. Two building blocks appear everywhere:

• Mass-action (A + B react): rate $= k\, x_A x_B$. Why? Reaction probability ∝ chance two molecules meet ∝ product of concentrations.
• Hill repression (protein P represses gene production): rate $= \dfrac{\beta}{1 + (P/K)^h}$. Why? As repressor $P$ rises, output falls; $h$ (cooperativity) sets how switch-like it is; $K$ is the half-max level.

Worked Example 1 — A negative feedback loop can oscillate (but needs ≥3 dimensions)

A gene makes protein $P$ that represses its own production; $P$ also degrades:

$\frac{dP}{dt} = \underbrace{\frac{\beta}{1+(P/K)^h}}_{\text{repressed production}} - \underbrace{\gamma P}_{\text{degradation}}$

• Why the first term? Hill repression: more $P$ ⇒ less new $P$.
• Why $-\gamma P$? Molecules decay at a rate proportional to how many exist.
• Crucial subtlety: this one-variable ($1$-D) autonomous ODE cannot oscillate. A single $\frac{dP}{dt}=f(P)$ can only move monotonically toward its steady state (where $f(P)=0$) — it never overshoots, because to overshoot $P$ would have to reverse direction, which requires another variable to "remember" the past.
• How to get oscillation: add memory. Either (a) an explicit time delay ($P(t)$ represses using $P(t-\tau)$), or (b) extra intermediate variables so the loop is at least $3$-D — e.g. the repressilator: gene A ⊣ gene B ⊣ gene C ⊣ gene A. Now the loop overshoots and undershoots → a self-sustained oscillation (a biological clock). No single molecule "oscillates"; the loop does. That is emergence.

Worked Example 2 — A toggle switch (bistability)

Two genes, each represses the other ($x$ ⊣ $y$, $y$ ⊣ $x$): $\frac{dx}{dt}=\frac{\beta}{1+(y/K)^h}-\gamma x, \qquad \frac{dy}{dt}=\frac{\beta}{1+(x/K)^h}-\gamma y$

• Why two coupled equations? Because the behaviour lives in the mutual interaction — remove the coupling and it's dead.
• Emergent result: the system has two stable states ("x-high" or "y-high") — a cellular memory switch underlying cell-fate decisions. Bistability appears only when $h>1$ (cooperativity). This is the mathematical heart of "the whole > sum of parts."

Worked Example 3 — Forecast-then-Verify

Common mistakes (Steel-manned)

Flashcards

Connections

• Reductionism vs Holism
• Gene Regulatory Networks
• Feedback Loops & Homeostasis
• Ordinary Differential Equations in Biology
• Hill Function & Cooperativity
• Bistability & Cell Fate Decisions
• Synthetic Biology (Toggle Switch, Repressilator)
• Dimensionality & Oscillations in Dynamical Systems

Concept Map

Hinglish (regional understanding)