Define systems biology and holistic modeling

6.5.1 · Biology › Systems Biology & Frontiers

YEH FIELD exist kyun karti hai?

Ek sadee tak biology reductionism se kaam karti rahi: organism ko tukdon mein tod do (genes, proteins, metabolites), har tukda akela study karo, publish karo. Isse humein parts catalogue mili (genome, proteome...).

Lekin har eent ko jaanna aapko ghar ki shape nahi batata. Dil ki dhadkan, ek circadian rhythm, ek oscillating gene circuit — yeh sab wiring se paida hone wale dynamic behaviours hain, kisi ek molecule se nahi.

SYSTEMS BIOLOGY kya hai?

Holistic model KAISE banate hain? (scratch se derive karo)

Chalo sabse seedha holistic model derive karte hain: interacting concentrations ka ek network.

Step 1 — State variables. Maano system mein $n$ components hain concentrations ke saath $x (t) = (x_{1}, x_{2}, \dots, x_{n}) .$ Yeh step kyun? Bina kya badalta hai yeh naam liye change model nahi kar sakte. Molecules ke liye concentration sabse natural variable hai.

Step 2 — Rate of change = production − consumption. Har component ke liye, mass balance kehta hai $\frac{d x _{i}}{d t} = (rate in) - (rate out) .$ Yeh step kyun? Yeh sirf matter ka conservation hai — pool ko sirf wahi cheez change karti hai jo enter ya exit karti hai.

Step 3 — Interactions rates ko doosre components par dependent banate hain. Ek network mein, $x_{i}$ ki production $x_{j}$ par depend kar sakti hai. Toh: $\frac{d x _{i}}{d t} = f_{i} (x_{1}, x_{2}, \dots, x_{n})$ Yeh step kyun? $f_{i}$ ka doosre variables se coupling exactly wahi hai jisko "network" kehte hain. Agar har $f_{i}$ sirf $x_{i}$ par depend karta, toh koi system nahi hota — sirf $n$ alag-alag problems hoti.

Step 4 — Concrete interaction terms. Har jagah do building blocks milte hain:

Mass-action (A + B react karte hain): rate $= k x_{A} x_{B}$ . Kyun? Reaction probability ∝ do molecules ke milne ka chance ∝ concentrations ka product.
Hill repression (protein P gene production ko repress karta hai): rate $= \frac{β}{1 + ( P / K ) ^{h}}$ . Kyun? Jaise repressor $P$ badhta hai, output girta hai; $h$ (cooperativity) batata hai yeh kitna switch-like hai; $K$ half-max level hai.

Worked Example 1 — Negative feedback loop oscillate kar sakta hai (lekin ≥3 dimensions chahiye)

Ek gene protein $P$ banata hai jo apni khud ki production ko repress karta hai; $P$ degrade bhi hota hai:

$\frac{d P}{d t} = repressed production \frac{β}{1 + ( P / K ) ^{h}} - degradation γ P$

Pehla term kyun? Hill repression: zyada $P$ ⇒ kam naya $P$ .
$- γ P$ kyun? Molecules us rate se decay karte hain jo kitne hain us par proportional hai.
Crucial subtlety: yeh ek-variable ( $1$ -D) autonomous ODE oscillate nahi kar sakta. Ek akela $\frac{d P}{d t} = f (P)$ sirf monotonically apne steady state ki taraf move kar sakta hai (jahan $f (P) = 0$ ) — yeh kabhi overshoot nahi karta, kyunki overshoot karne ke liye $P$ ko direction reverse karni padegi, jiske liye ek aur variable chahiye jo past "yaad" rakhe.
Oscillation kaise paayen: memory add karo. Ya toh (a) ek explicit time delay ( $P (t)$ repression karta hai $P (t - τ)$ use karke), ya (b) extra intermediate variables taaki loop kam se kam $3$ -D ho — jaise repressilator: gene A ⊣ gene B ⊣ gene C ⊣ gene A. Ab loop overshoot aur undershoot karta hai → ek self-sustained oscillation (ek biological clock). Koi single molecule "oscillate" nahi karta; loop karta hai. Yahi emergence hai.

Worked Example 2 — Toggle switch (bistability)

Do genes, har ek doosre ko repress karta hai ( $x$ ⊣ $y$ , $y$ ⊣ $x$ ): $\frac{d x}{d t} = \frac{β}{1 + ( y / K ) ^{h}} - γ x, \frac{d y}{d t} = \frac{β}{1 + ( x / K ) ^{h}} - γ y$

Do coupled equations kyun? Kyunki behaviour mutual interaction mein rehta hai — coupling hatao aur sab khatam.
Emergent result: system ke do stable states hain ("x-high" ya "y-high") — ek cellular memory switch jo cell-fate decisions ke peeche hai. Bistability tab hi aata hai jab $h > 1$ (cooperativity) ho. Yahi "whole > sum of parts" ka mathematical core hai.

Worked Example 3 — Forecast-then-Verify

Worked example Compute karne se pehle predict karo

3-gene repressilator lo (A ⊣ B ⊣ C ⊣ A) strong repression ke saath ( $h$ bada). Forecast: loop settle nahi ho sakta — har gene pichle wale ko chase karta hai, toh concentrations ek doosre ko circle mein chase karte hain ⇒ sustained oscillation. Verify: 3 coupled ODEs simulate karo; dekho A, B, C ek repeating relay mein rise aur fall karte hain. Ab steel-man: agar aap circuit ko ek single self-repressing gene (1-D) par collapse kar do, wahi strong cooperativity aapko ek sharp switch lekin bilkul bhi oscillation nahi dega — ek 1-D autonomous ODE hamesha monotonically apne steady state ki taraf relax karta hai. Sabak: oscillation ke liye cooperativity aur enough coupled components / ek delay dono chahiye; parts list akele yeh kabhi reveal nahi karti.

Figure — Define systems biology and holistic modeling

Common mistakes (Steel-manned)

Flashcards

Systems biology ko reductionism se kaunsi core idea alag karti hai?

Yeh components ko ek interacting network ke roop mein study karta hai aur emergent, system-level behaviour par focus karta hai, isolated parts par nahi.

Emergent property define karo.

Ek property jo poore system mein present ho lekin kisi bhi isolated component mein absent ho (jaise oscillation, bistability, homeostasis).

Holistic dynamic model ki general form kya hai?

\frac{d x}{d t} = f (x)

— har component ki rate doosron par depend karti hai.

\frac{d x _{i}}{d t} = f_{i} (x_{1}, \dots, x_{n})

ek "network" kaise capture karta hai?

Kyunki

f_{i}

doosre variables par depend karta hai; equations ke beech coupling HI network hai.

Mass-action term

k x_{A} x_{B}

kya represent karta hai, aur product kyun?

A+B ke react karne ki rate; product isliye kyunki collision probability ∝ do molecules ke milne ka chance.

Hill repression term likho aur batao

h

kya control karta hai.

\frac{β}{1 + ( P / K ) ^{h}}

;

h

= cooperativity, control karta hai repression kitna switch-like (sharp) hai.

Kya ek single self-repressing gene (1-D ODE) oscillate kar sakta hai? Kyun/kyun nahi?

Nahi — ek 1-D autonomous ODE mein koi memory nahi aur yeh monotonically

f (P) = 0

tak relax karta hai. Oscillation ke liye ≥2 state variables ya ek time delay chahiye.

Genetic oscillator ke liye minimal ingredients kya hain?

Negative feedback + enough dimensions (≥3-gene loop jaise repressilator, ya ek explicit time delay) + sufficient nonlinearity/cooperativity.

Mutual repression of two genes se kaunsa emergent behaviour aata hai?

Bistability — do stable states wala ek toggle/memory switch (chahiye

h > 1

Steel-man: "sare genes jaanna ⇒ cell samajhna" fail kyun hota hai?

Behaviour interactions mein rehta hai; identical genomes alag-alag stable states mein ho sakte hain.

Kya emergence mystical hai?

Nahi — yeh nonlinear coupled ODEs ka feedback aur enough dimensions ke saath mathematical consequence hai.

Recall Feynman: ek 12-saal ke bachche ko explain karo

Socho teen bachche ek circle mein hain, har ek ko bola gaya hai ki jo apne pehle wala loud ho jaye usse quiet karo. Koi kabhi settle nahi kar sakta: pehla quiet ho jata hai, toh doosra loud ho jata hai, toh teesra quiet ho jata hai, toh pehla phir loud ho jata hai — ghumta rehta hai hamesha. Yeh never-ending relay ek clock hai, aur yeh sirf isliye exist karta hai kyunki teen hain looping mein. Agar sirf ek bachcha hota jo khud ko quiet karne ko kaha jaata, woh simply ek aaram deh medium volume tak pahunch jaata aur wahan reh jaata — koi rhythm nahi. Systems biology saare bachon ko aur unke rules ko saath mein study karta hai, kyunki dance connections aur enough players se aata hai, kisi ek bachche se nahi.

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