Measuring Complexity in Biological Systems

[ryokan]

The term "complexity" is currently used in the study of non-linear dynamics.
The main problem with this term arises when is applied to biological systems.
For example, does evolution generate ever more complex organisms?
A good measure of complexity is lacking. We can observe complexity at structural, functional or "grammatical" levels, seeing the number of specialized cells in one organism, its functions and the nesting of functions, or some relationship between genome sequence and structure, function or, more generally, fitness to environment.
The question remains: How could we measure complexity in biological systems?

 

[selfAdjoint]

The term "complexity" is currently used in the study of non-linear dynamics.
The main problem with this term arises when is applied to biological systems.
For example, does evolution generate ever more complex organisms?
A good measure of complexity is lacking. We can observe complexity at structural, functional or "grammatical" levels, seeing the number of specialized cells in one organism, its functions and the nesting of functions, or some relationship between genome sequence and structure, function or, more generally, fitness to environment.
The question remains: How could we measure complexity in biological systems?

A discussion of this issue, with a notice of how not to do it (as exhibited by the ID people) is at http://cscs.umich.edu/~crshalizi/weblog/000217.html.

 

[ryokan]

A discussion of this issue, with a notice of how not to do it (as exhibited by the ID people) is at http://cscs.umich.edu/~crshalizi/weblog/000217.html.

Thank you for this link.
Nevertheless, that surprises me. I am not "ID people". I have only suggested a thread that I think attractive.
Complex systems are interesting. Some Nature's papers and insights have been devoted to this topic. To pose questions on complexity don't implies creationism or ID viewpoints, as you seem suggest to me with your link.

 

[ryokan]

Are the complexities in artificial (actual and potential future engines) and biological systems comparable? How?

 

[Phobos]

For example, does evolution generate ever more complex organisms?

No. There are examples of some organisms like parasites that have become simpler.

Evolution need not be "improvement"...only "change".

 

[iansmith]

No. There are examples of some organisms like parasites that have become simpler.

Simpler might not be the best term. More specialize would be more appropriate. A parasite has become so well adapted to its environment, it cannot do without it.

For bacteria, there such a thing as genome reduction. You can have an Streptomyces, around 9 million bp genome, and E. coli, 4 to 5 million bp, compare to an obligate parasites or endosymbions such as Haemophilus influenzae, between 1.5 to 2 millions bp, Chlamydia, around 1 million bp and mycoplasma and Wigglesworthia glossinidia endosymbiont of Glossina brevipalpis, both at less than 1 million bp.

Complexity is also in the eye of the beholder. Are the mechanism that complex or just a juxtaposition of simple elements?

 

[Phobos]

Thanks for the clarification, iansmith. I agree.

Another answer to the question "does evolution generate ever more complex organisms?" is to note that the majority of life, now and in the past, is bacterial. "Complex" multicellular organisms are practically a fringe situation to the main aspect of life.

 

[ryokan]

Thanks for the clarification, iansmith. I agree.

Another answer to the question "does evolution generate ever more complex organisms?" is to note that the majority of life, now and in the past, is bacterial. "Complex" multicellular organisms are practically a fringe situation to the main aspect of life.

I doubt. How many bacterial species are there? How many insects?

 

[NanoTech]

Could we associate the evolution of complexity of environment to morphological complexities in organisms. Did complex behaviors/physical traits evolve with the environment, or with random mutations in DNA?

 

[iansmith]

I doubt. How many bacterial species are there? How many insects?

This question was answered in the following thread
https://www.physicsforums.com/showthread.php?t=54142

the discussion was going off topic.

 

[ryokan]

This question was answered in the following thread
https://www.physicsforums.com/showthread.php?t=54142

the discussion was going off topic.

Yes. I agree. One thing is diversity and another is complexity, although both be related.
I suggest to center the discussion on a level: one eukaryotic cell, an hepatocyte for example. How could we compare its complexity with that of other cell or an electronic chip?
Would it make sense here the use of measures of complexity arising from other fields (Shannon's entropy, algorithmic complexity and so on) ?

 

[loseyourname]

How would you compare entropy between an electronic chip and a eukaryotic cell? The measure of free energy would increase with every new bond, and as such, could increase simply because one contains more molecules than the other. Greater volume hardly seems to suggest additional complexity.

I would propose that the best measure of complexity in a cell is the number of functions that cell can serve. For instance, an embryonic stem cell, because it is in theory capable of carrying out any function that a human cell can carry out, is more complex than a fully differentiated neuron, which can only perform neuronal functions exclusively.

 

[ryokan]

How would you compare entropy between an electronic chip and a eukaryotic cell? The measure of free energy would increase with every new bond, and as such, could increase simply because one contains more molecules than the other. Greater volume hardly seems to suggest additional complexity.

I would propose that the best measure of complexity in a cell is the number of functions that cell can serve. For instance, an embryonic stem cell, because it is in theory capable of carrying out any function that a human cell can carry out, is more complex than a fully differentiated neuron, which can only perform neuronal functions exclusively.

I didn't refer to thermodynamic entropy, but to Shannon's approach to measure of information.
I agree that the number of functions by cell could be a good measure, but if so, we have difficulties with the term function. In your example, an embryonic stem cell is capable of carrying a lot of functions, but only capable. There are potential and real functions. From a real, non potential viewpoint, a neuron could be more complex. I don't know.
How could we define a function? Reproduction is a function, but involving a lot of functions at different levels, molecular, substructural...
The study of gene expression with biochips could light the question, but it remains the number of interactions among all the molecules with a role in gene expression.
Furthermore, It is possible to study complexity at qualitative (type of function and molecular interactions involved) or quantitative (number of genes expressed) levels.
What do you think?

 

[loseyourname]

We could always use the simple mechanical view of complexity: The number of moving parts that perform work. This might be a little more representative than the number of functions performed, because you could easily argue that a machine that uses more parts to perform the same function as machine with less parts is more complex. Might be difficult to quantify on a cellular level, though.

 

[ryokan]

We could always use the simple mechanical view of complexity: The number of moving parts that perform work.

If so, the number of different molecules by cell would be a measure of complexity. Yes?

you could easily argue that a machine that uses more parts to perform the same function as machine with less parts is more complex.

I doubt. It is possible that more parts for a same function reflect only redundant effects or be a side effect of evolution (gene duplications, for example)

 

[loseyourname]

If so, the number of different molecules by cell would be a measure of complexity. Yes?

Not necessarily. For example, additional structural proteins (say, just because one cell is larger than the other), wouldn't qualify as additional complexity. But a glycoprotein would be more complex than a simple protein. Compound molecules, in this way, are more complicated. And of course, motor proteins, and flagella, and other moving molecules that actually form elementary machines, are more complex than stationary molecules.

I doubt. It is possible that more parts for a same function reflect only redundant effects or be a side effect of evolution (gene duplications, for example)

Redundancy is a form of complexity, though. Wouldn't you consider a Rube Goldberg device that turns on a light more complex than a simple switch. It may not be any more advanced or effective, but it's certainly more complex. In the same way, the cell wall of a gram positive bacterium is more complex than the cell wall of a plant, though both serve essentially the same function equally well. Dolphin song is more complex than the human voice, but humans are still capable of conveying far more information and carrying out more advanced levels of communication. Of course, this is because human language is more complex, but there is still no question as to the relative complexities of respective vocal apparati.

 

[ryokan]

Not necessarily. For example, additional structural proteins (say, just because one cell is larger than the other), wouldn't qualify as additional complexity. But a glycoprotein would be more complex than a simple protein. Compound molecules, in this way, are more complicated. And of course, motor proteins, and flagella, and other moving molecules that actually form elementary machines, are more complex than stationary molecules.

Then, we pass from complexity based in number of distinct molecules to complexity based in types of molecules. I agree. Glycoproteins are very complicated. At molecular level then, how could we measure complexity? Here, one possible approach would be the use in some form, of algorithmic complexity.
Alternatively there is the possibility to use periodical structures into a molecule. DNA would be so more "simple" than a glycoprotein (in terms of description). Dynamic changes, as you suggest, would increase the grade of complexity, even for one molecule such as an enzyme interacting with its substrate with a subtle tridimensional change.

 

[ryokan]

Dolphin song is more complex than the human voice

I don't understand this assert, but I find it very interesting.



How is compared the complexity of human and dolphins songs? In terms of frequency spectrum?