1 of 29

Modularity as a dynamic and functional phenomenon

Herbert Sauro, Steve Andrews, Joe Hellerstein: University of Washington, Seattle,

Song Feng, Steve Wiley: PNNL, WA

2 of 29

Terminology

There is some terminology worth mentioning:

Mechanism (or subsystem)

Two or more biochemical processes forming a network where species evolve in time (John J Tyson, Katherine C Chen, Bela Novak, Sniffers, buzzers, toggles and blinkers: dynamics of regulatory and signaling pathways in the cell, Current Opinion in Cell Biology, 2006)

Motif

A topological pattern that is found enriched in interaction graphs (Milo R, Shen-Orr S, Itzkovitz S, Kashtan N, Chklovskii D, Alon U: Network motifs: simple building blocks of complex networks, PNAS, 2002)

Module

To be discussed, but there are many interpretations (18+ on Wilipedia’s page)

3 of 29

Terminology

Mofits

Subsystems

Modules

Tyson, Chen, Novak

Mangan & Alon

?

4 of 29

Origins

The modern use of the word module* started in the 1950s with the entry of the word into the dictionary to mean “interchangeable part”.

New post-war electronics companies were building modular systems where faulty “modules” could be pulled out and easily replaced. NASA also picked up the term by to refering to the "lunar module" as a separate self-contained system in 1961.

PDP-8 transistor/diode logic module (triple flip flop, circa 1965/66)

The use of the word module grew in in biology in the late 1990s.

https://www.computerhistory.org/revolution/minicomputers/11/331/1905

*Originates from the Latin modulus, meaning "a small measure".

5 of 29

Trends: Reported in Books

books.google.com/ngrams/

6 of 29

In Biology

In biology the word module came to promince with the publication:

“From molecular to modular cell biology”, Leland H. Hartwell, John J. Hopfield, Stanislas Leibler and Andrew W. Murray, 1999.

Hartwell et al emphasized the functional aspects of a module:

“A functional module is, by definition, a discrete entity whose function is separable from those of other modules.

Since then the use of the word module has dramatically increased. I did a PubMed search for the word ‘module’ and it picked up:

80,402 matches

7 of 29

On the other hand:

In the Encyclopedia of Systems Biology, the chapter on modules defines it thus:

“A biological module is a subnetwork entity composed of nodes that are more closely related among them than to the rest of the components in a network;”

Application of Graph Theory for Identifying Connectivity Patterns in Human Brain Networks: A Systematic Review - Scientific Figure on ResearchGate. Available from: https://www.researchgate.net/figure/of-global-graph-measures-A-Segregation-measures-include-clustering-coefficient-which_fig4_333968671 [accessed 12 Apr 2026]

The notion of function has disappeared

Modularity and community structure in networks

M. E. J. Newman, PNAS, 2006

�The structure and function of complex networks, 2003

https://arxiv.org/abs/cond-mat/0303516

SIAM Review

8 of 29

Other uses of the word module in biology

Some common occurrences of the word module include:

  1. Operons, especially in bacteria.
  2. Developmental modules such as systems responsible for development of specific parts like limbs (Waddington first proposed this idea in his 1957 book strategy of the the genes but didn’t call them modules). E.g conserved transcritional networks in Sea Urchins.
  3. Protein complexes that perform a specific function
  4. Protein interactions maps that show densely connected nodes separated by low density connections.
  5. Networks with some specific behavior, similar to the Tyson networks.
  6. Anything in synthetic biology

9 of 29

What might be the general consensus?

The word module seems to be attached to anything that is a collection of closely connected parts, be it genes, biochemical neworks, graphs, protein subunits etc. Is that a sufficient definition? Is this the question for ths week?

Copilot generated.

10 of 29

Waddington’s Famous Epigentic Landscape

Genes influence the modular landscape by which organisms develop.

The Strategy of the Genes: A Discussion of Some Aspects of Theoretical Biology,

Waddington, C. H. [ With an Appendix by H. Kacser, Pd. D. ], 1957

11 of 29

Returning to the 1950s. What characteristics would an engineer of the 1950s list as particular to a module?

  1. A defined Interface
  2. A defined function (flip-flop)

12 of 29

Returning to the 1950/60s. What characteristics would an engineer of the 1950s list as particular to a module?

  1. The outputs have very low output impedance, meaning that increases in load have little effect on the output generated by the module.
  2. In constrast, inputs had a very high impedance, meaning inputs will not heavily load the output from another circuit.
  3. The last two properties ensure that a module retains functional integrity when connected to other circuits. This is what makes engineering complex systems possible.

13 of 29

Summary Characteristics of a Module

We can summarise the characteristics of a hardware module in the most abstract sense as:

Well-defined interface: ports are explicitly declared and interactions are constrained to those ports

Well-defined function: the input-output behaviour is specified and predictable

Composable: can be connected to other modules without side effects

Biology is going to be a bit more messy that this but these are probably the basic characteristics. There is no encapsulation because we’re in a liquid phase but it makes it easier to change the interface.

The interface defines the boundary.

The function characterises what crosses that boundary.

Composability is the guarantee that when modules are assembled, the functional characteristics are guaranteed.

14 of 29

Summary Characteristics of a Module

Function

Interface

Composability

15 of 29

Modularity can also be hierarchical

Human engineers use this strategy all the time

16 of 29

Taking an Engineering Approach to Biology

This idea was taken up by Domitilla Del Vecchio at MIT. She drew a parallel between what electrical engineers do and what biology might do.

17 of 29

Available Mathematical Approaches

  1. Dynamical Systems Theory

Classical Differential Equation Analysis

Focuses on stability, bifurcation analysis and chaotic dynamics, role of positive and negative feedback in stability.

  1. Martin Feinberg's Chemical Reaction Network Theory (CRNT) et al
  2. Bruce Clarke’s Stoichiometric Network Analysis (SNA)

Both Feinberg and Clarke focus on stability analysis based on topology alone

  1. Control theory and Biochemical Control Theory

Mixture of stability, sensitivity (robustness) and the role of feedback , primarily used for design in engineering and reverse engineering in biology.

Copilot generated.

18 of 29

Introduction to Biochemical Control Theory

Control theory is a good choice for understanding biochemical networks and modularity. It directly addresses:

  1. Feedback patterns
  2. Robustness
  3. Stability
  4. Is able to relate system behavior to the component parts, i.e is explains phenotype in terms of genotype and environment.

19 of 29

Introduction to Biochemical Control Theory

Later in the week we will be giving an instructory tutorial on biochemical control theory as well as new work we have done on how to define a modularity coefficient given a particular ‘module’ and its context.

You’ll also hear more on this topic from Joe Hellerstein, Steve Wiley and Steve Andrews and ways in which modularity can be defined.

Copilot generated.

20 of 29

Acknowledgements

BIC: Evolving Signal Processing Circuits from Biological Reaction Networks

Award Number: 0432190, 2004

Extension of Metabolic Control Analysis and Biochemical Systems Theory to Stochastic Systems

Award Number: 0827592, 2009

Reinhart Laubenbacher’s monthly modularity meetings��Extensive discussions with Steve Andrews, Song Feng, Joe Hellerstein, Steve Wiley

21 of 29

Resources

Online book (under construction)

https://hsauro.github.io/mcabook/index.html

Actual books:

A Frequency Domain Approach to Sensitivity Analysis of Biochemical Networks�Brian P. Ingalls�The Journal of Physical Chemistry B 2004 108 (3), 1143-1152�DOI: 10.1021/jp036567u

22 of 29

What do we mean by retroactviity?

To the left is a TTL NAND gate chip, part of the 7400 series. Developed in the mid 1960s, it revolutionized digital electronics for professionals and hobbyists.

It offered two two key benefits: 1) all chips in the series worked off a standard 5v rail. 2) Secondly an output could connect to up to 10 other inputs without signal degradation.

23 of 29

From these may cases what can we distill out?

1) Interchangeability/reusability, �2) Function independence, �3) Defined interface, �4a) Evolability�4b) Modifiability, �5) Standardization�6) Structural separability (different from functional independence, eg a NAND gate is functionally separate but a network graph is structurally separable since it has no intrinsic function);

24 of 29

Generating functional networks using evolution

This network will compute the square root of the input x. The square root will be the concentration at node 3

25 of 29

Here is a range of networks that can compute the square root:

26 of 29

Solve quadratic equation

27 of 29

Cube Root Calculator

28 of 29

Other Evolved networks:

  1. Bistable switches
  2. Oscillators
  3. Low pass filters
  4. High pass filters
  5. Band pass filters
  6. Homeostatic systems

29 of 29

But there are many other definitions of modularity

The word modularity is one of the overused words and to some extend has lost its meaning.