Caution: Mild geek content
A lazy weekend afternoon at home. I'm watching TV with a roommate and he makes a comment on seeing yet another ad for a low-calorie food product, "The American view on diet is so f***ed up, man."
I smile, and taking one feather from my own cap and one from my Mom's, I dive into an explanation of how the American ideas on diet are reductionist in nature as opposed to the traditional Indian view. We never said ghee is bad for health just because it has hydrogenated fat; we know that there is little harm in having some ghee with rice as long as vegetables and pulses form the greater proportion. I stopped myself in the middle of a sentence when I realized that I was lecturing - I seem to be doing a lot these days. Perhaps this PhD is turning me into a professor, whether I want to be one or not!
But the funny thing is that I'd never voiced that particular point of view before; so where did it come from? The answer: my research. I think I started seeing the distinction between Reductionist and Holistic more clearly after I read a book on Chaos theory.
Chaos theory attempts to explain things where conventional science fails. It says that there are some systems where the equations or rules simply cannot be solved because of the interdependencies; and they cannot be simulated beyond a point because they are extremely sensitive to the variables. There is a famous story in which Edward Lorenz was simulating a weather system, and noticed a re-run of the simulation diverging completely from the original run even though it was the same system with the same initial conditions. So, he re-checked the initial conditions and realized that he'd rounded the floating point decimals down to a precision of six (I think) digits. Essentially, a difference of 0.000001 had yielded a completely different result after just a few iterations!
Think about the fact that in real life, there could be a million different external or internal perturbations on a variable, not to mention the difficulty in measuring values accurately beyond a certain point. Chaos theory accepts the infeasibility of studying such systems in the reductionist manner, and instead tries to identify patterns. Today there is quite a bit of formal mathematics involved, but initially, chaos theorists were looked upon as quacks.
My next leap in understanding came when I studied complex networks. This is the science of naturally evolved, large scale networks in a plethora of fields, covering everything from Social Networks to The Internet to Cellular Biology. Why, if you consider co-stars in a movie as linked, then on considering all Hollywood movies ever made, you will come up with a huge actor collaboration network. The science of complex networks identifies amazing similarities in all these networks, and the fact that such a complicated interaction network can yield results that are extremely counter-intuitive. Again, you cannot simply break down the system to the individual parts and rules and expect to easily predict the behavior of the system. With a multi-layered, large-scale network, taking the reductionist approach is but the first step in understanding the system.
Systems Biology is the science of life studied in a holistic manner. As one scientist put it, we are trying to put Humpty Dumpty back together again. We have decoded the entire human genome; yet, we are far away from understanding how the same genes work together to create proteins, which work together to regulate individual biological processes, which work together to "run" a living organism. The same kind of problem is being faced in many, many other fields.
Chaos theory and non-linear dynamics, complex networks and systems engineering; these are are all interconnected fields that try to understand things - and perhaps build systems, with their inspiration - from a holistic viewpoint. These are the new challenges in Science; Reductionism is almost passé, and is increasingly being thought of as the first step in a long process. It is extremely humbling - and interesting - for a scientist to study and understand such complexity in nature. Our ancestors were nearly as good at it as we are now!
Watch your diet,
Cheers,
Prashanth.
Postscript: I made a start to the Science Blog that I was talking about, but decided that I would just include a few semi-scientific posts in this blog. This is one of them.
Subscribe to:
Post Comments (Atom)
9 comments:
mild geek content?!?...it takes time to be careful about what you eat, but it is important...its a bit harder to do when traveling :)...
Intern,
Milk gild content? Now thats a new one :D
Well, I just meant to say that we in India have the holistic approach to food by tradition. It's not the same in the US, there are more misconceptions than truths among the common folk.
Alraqs,
Oh yes, defy just mild :) wait till I really get into gear! Hope you're having fun with your travels... am expecting some nice blog posts on that soon!
interesting..
i thought systems biology seeks to be reductionist in an area where things are holistic..
you know what the heart does but you don't know how exactly it works...
now..you try to figure out its mechanisms at a lower level..
Nice Post Prashanth. I suspect the Holistic approach is generally far more nontrivial to analyse .... thus most work has been using reductionist approach....
@Prashanth @Ramani: An holistic approach of the cell is, as u guys know, one of the most major lines of invesitgation that biology would probably like to achieve.. Problem is an holistic approach of the cell is never going to be easy... I think molecular biology was good as a starting point to understand a lot of aspects of the molecules that make up a cell, but at this point of time, it is time to go build up the studies to go beyond single molecule chemistry... I think it is a very good point to build up a whole metabolic pathway in wet lab chemistry and to understand the chemistry of these enzymes when they come together. At that point, systems biology can achieve more realistic models of a whole metabolic pathway, for example. Secondly, in a cell, nothing is in isolation. There are all interconnected. Its important to know which aspects of the cell can complete a layer. For example, a cell could form a system for study but one should realize that the inputs to this cell should be chemicals that could be a function of the environment. The output of one biological cell is the input of another cell. In other words, you can model a cell, but at the same time, you should realize the limitations of the model because no cell can be found in isolation and even microbes are found in communities - the community is made up not of a homogeneous population of cells (unless it is a lab experiment) but there are various species of microbial cells that make up the community. The problem with systems biology right now (and why biologists hate it) is because:
a) very little of the atomistic/molecular details can be understood from it.
b) Very little experiments can prove or disprove a model and most of these guys do not collaborate a lot to do the above.
c) Biologists know that since the knowledge of each individual protein or different proteins is incomplete, these models have to be incomplete in nature and hence, they think that systems biology theoreticians can prove anything.
Ramani,
Discovering those lower level functions are by themselves not systems biology, just the first step. Its discovering the inter-relationships in putting them back to get the whole picture that is really termed as systems biology.
sd,
Thanks... well, all science is based on foundations of axioms and theorems and such, so science is inherently reductionist. Unfortunately it doesn't always work :)
B-a-l,
You're right on most counts but I disagree with one point... the atomistic/molecular details are necessary for building a systems biology model. They may not be immediately apparent on inspection, but they are definitely in there. But it is true that the incomplete information results in incomplete models... and yet, they must start somewhere!
@prashanth: Dont get me wrong... I do believe that some of the details are put in... What I meant with that part of the comment is explained better now: Some good theoreticians (the theoretical equivalent of a molecular biologist) can do molecular dynamics simulations of proteins with their substrates and understand certain aspects of the chemistry. To understand the chemistry completely, we need a better way to integrate QM and molecular dynamics and that is certainly a goal for the future... These kind of simulations try to model the atomistic details of how a biomolecule functions.... However, when you do systems biology approach, one can not perform a MD simulation (with current day computer power) with atomistic detail because of the enormous amount of atoms in the system... So, one makes a coarse grained model of the system. For example, in the cellular automata kind of model - the simulation takes place in a lattice. Each cell in the lattice has a state decided by what molecules are present on it. There are certain rules for diffusion of molecules and certain probabilities of reaction occuring, etc. These kind of models are being developed a lot in systems biology but they by themselves do not explain how each molecule functions - why does the enzyme catalyze the reaction, etc. The probabilities of the reaction occuring is decided by the rate of catalysis but not the atomistic details of why it happens. It can be argued whether you really need an atomistic level description of the whole cell but it certainly has been thought about by nature. The coarse grained model does not explain why a molecule functions the way it does but they do explain what happens when these molecules (with their appropriate rate constants and their appropriate diffusion rate and all the intermediate states governed by the model) come together.... If you find some systems biology approach that explains the rates of the molecule and does not use it as a parameter, please tell me about it.... I might be wrong or limited in what I have read...
BTW, a book you might like to read on trying to take a holistic approach of the cell (without being too technical) is The way of the cell by Francis Harold.
Ok, I get you now. Naturally you are more knowledgeable than me on these specifics, but I get the feeling that the debate you are talking about is more about 100% causal modeling vs. some abstraction, rather than conventional approach vs. systems approach. I think the systems biologists are getting flak for over-abstraction, but they can't help it because of incomplete information or lack of processing power.
Thanks for the ref, will defy check it out.
Post a Comment