Slashdot Mirror
Searching for Life's Blueprints
Makarand writes "If the
claims made by the accomplished biophysicist Andras Pellionisz hold any water, life's blueprints may indeed be in fractal patterns found in the DNA.
In a human, genes constitute only around 2-3% of the total DNA (the exons). The rest
of the non-genic DNA (called introns) play a role that has not
yet been understood and some have even suggested that these
may merely be evolutionary leftovers. Removal of this "junk-DNA", however, has been proven to be lethal. The introns, he claims, may have
the "building construction blueprints" in the form of fractal patterns that
the exons use to build living tissue. A patent application covering attempts to count, measure and compare the fractal properties of introns for diagnostic and therapeutic purposes has been made.
He hopes his patent will help him launch his company and make it a key player in this field."
Perhaps you might say DNA is code, and proteins are objects? I think DNA is like a C header file, really -- it specifies the structural information, but leaves out the mechanics, which come from physics.
In any event, the mechanisms by which DNA is used to create proteins are actually very well understood already. Understanding what proteins do after they're created -- i.e., fold up, catalyze chemical reations -- now that's another story. But that doesn't mean we don't understand how DNA is used to create them.
* Well, the purpose of some proteins is to transcribe DNA and thereby build other proteins, but that's not what most proteins do.
Cantankerous old coot since 1957.
Because genes are so neatly digital, people tend to think of them as being all the information in a cell. Actually a cell is like a computer - some information is stored explicitly as software, and the rest is hardware, but it is really hard to tell where software ends and hardware begins. After all, floating point coprocessors work with big internal tables - so are those tables hardware or software? And think of the IBM keyboard BIOS. It started out life as hardware and ended up as software that emulate the hardware. Similarly, some proteins are taken from the same gene, but with a different set of exons.
Another thing is that biological systems love hacks, so the borders are constantly shifting. I think there's a good chance of finding introns doing unexpected stuff.
I bragged about my Karma at a job interview but I didn't get the job.
first it's well known that DNA is not merely a double helix but this ribbon also coils on itself (super coiling) and can be would in complex patterens around the biological equivalent of tape reals (called histones). And that there even larger hierarchies of organization like chromosomes.
When a gene is "expresses" (read) from the DNA, that portion of the DNA has to be exposed, thus from square one the mobility and ease of exposure of a structure regulates its expression. Additionally, in order for some of the portien moelcules that trigger expression as well as those that do the expressing to bind to the DNA the DNA often has to have a characterisitic kink or lack of a kink. Binding in biology is --unlink the interaction of simpler molecules--inherently recognition of another structure.
so point 1 is that whoop-tee-do structure of DNA organization is important to expression. We all knew that already.
The second point is that as far as binding goes these specific events are almost excusively local. that is proteins and other molecules that bind to DNA are small (relative to the size of DNA), sort of like a fly landing on an aircraft carrier. At the scale of the dimensions of binding we are takling about atomic interactions and as the word "atomic" suggests, there is no notion of fractal subdivsion of space available. In other words patterns that exist distantly elsewhere in the DNA have no relevance to a binding event.
The third point to make is that the are many useful properties of "useless" sections of DNA. For example, at various times in its lfe DNA breaks the double helic and becomes two complimentary strands over sections of the DNA. Sometimes the one strand from won pair will go bind with a strand from another pair. This mainly happens when the two strand-swapping sections of DNA have nearly comlimentary chemical (base or nucleotide) patterns. At this chemical interaction level, whether or not the DNA section in question is "codeing" (and exon) or non-coding (an intron) is moot. DNA is DNA. thus non-coding regions can facilitate strand pairing and strand swapping activity. In other words useless DNA has a purpose of structure-structure interaction. TO the extent that this is already known this patent issue is silly.
Now What about those introns are they really useless DNA? some may be, some are not. Its a little tricky to exaplain in a few words but you have to imagine DNA like a hard disk with streams of consecutive bits. the word size of reading this is 3 bits. however, one has a slight problem when you go to read it, where do you start reading? if you are off by one bit then each word contains 2 bits from one word and one bit from the next word. this is called a frame shift, and obviously there are three possible frames on could read words in. Amazingly enough, not only can the cell figure out which frame to read, but sometimes all three frames contain a valid message!!! its a lot like the winnowing and chaffing encryption scheme. (indeed sometimes the messages can be read backwards and in a different frame to make sense too, much like a palindromic sentence, except that the reverse sentence may be different but still make sense). One purpose of introns is to create frame buffers and other signals to guide the readin mechaism to get into the proper frame.
Another purpose of introns is what is known as alternate splicing. Sometimes as (or after) a message is read off, sections of the dna get skipped over, like jumping a track on a vinyl record, and discontiguous portions of a the message are joined together. The decision to skip or not to skip can be regulated. Thus he same nominal section of DNA can produce slightly different edited messages. Thus introns sort of multiply the number of gene variations.
Finally, because of the way DNA makes mistakes when it copies it self or repairs damage, what offen happens is that a chunk of DNA gets copied to a new place on the DNA and the old one is not completely erased. This is infact exactly like a fragmented hard disk. Image a hard disk in which you have copied the smae files many times, and deleted the ones. At this point the FAT table fets lost and you have to use norton disk recover to try to find files. Wll you find lots of complete files and also fragments that look like old versions of parts of other files. This is what DNA looks like. So these self-similar patterns actually emerge accidentally. Since the chunk size varies the sel-similar patterns can be multi-scale and hence are fractal like. This is all accidental! Now its possile to imagine that what was once accidental is now being exploited by the body for a new puprose. For example, recombination plays a role in the immune system. But I doubt that the fractal nature of this is important. One reason to doubt it is that it is simpler to imagine that this happens beacuse there is no penalty for it happening. In higher organisms having wad's of extra DNA does not harm the cell since higher orgnaism have lots and lots of error correcting mechanisms to deal with DNA damage, dealing with extra DNA is small potatoes. Conversely, single cell organisms have a preimum on efficiency and thus minimize the saize of their DNA. Bacteria for example dont have introns, and have very little junk DNA. Viruses almost never have any junk dna at all bacause space is at a premium. Thus biology shows that when there is a reaosn to do so organisms chuck extra DNA.
so in conclusion I think this idea is cute but really nothing new or special, and is probably mostly hokum.
Some drink at the fountain of knowledge. Others just gargle.
I used to think that science was the last field which blatant greed had not infested yet, and I am proven wrong yet again...
Yeah...academics used to do their thing for the massive ego gratification, now they do it for profit. Don't know that it's necessarily worse this way.
It's not like scientists were ever this pure group of unbiased, purely objective people who are solely out to benefit the world and increase the knowledge of all. That's the publicity answer. Fact is, we fall to the same weaknesses as everyone else, including the great god Profit, and this shouldn't be surprising.
For what it's worth, the worst example was of a couple of guys, Ziegler and Natta, who invented a class of catalysts while working at a university. They worked really well, so they left the university (who paid for the research) and started a company, without giving the university a dime. They made millions, I believe. It happened in the 50's. So this isn't really new...though more widespread as universities have realized they can make a lot of money that way (patents) without much effort.
-Looking for a job as a materials chemist or multivariat
Protiens are what turn DNA into other proteins. Those new and existing proteins then interact with eachother, so the VM/bytecode analogy works better than you think. I would suggest that the protiens might be analogus to instantiated objects. They intract with eachother in a manner that is not necesarily dependant on the underlying bytecode or VM.
Of course, the idea of the virtual machine isn't necessary, since the machine is real.
DNA is less like a C header file than it is like a C program.
The basic mechanisms by which DNA is used create proteins is well understood on one level. What isn't well understood is all the regulatory mechanisms that interact with that process.
Why is a given stretch of DNA transcribed into mRNA? Why is it that only particular pieces of a stretch of DNA end up in the mRNA. What determines how long that mRNA transcript exists in a cell, and how many times it is translated into a protein?
These are all open questions before you even get to the question of protein folding and catalysis, and you have to assume that all the proteins in a cell have some involvement in gene regulation. How does a muscle cell know that it is time to stop producing myosin. Somehow, there is a feedback loop.
Not to nitpick but the post is wrong about introns. Introns are actually transcribed into RNA but then spliced before export. Most of the DNA isn't even transcribed into RNA, thus the overwhelming majority of the DNA is not made of introns.
The mechanisms by which genes code for structure
are reasonably well understood. There's no need
to invoke mysterious fractal magic to explain
it.
A good non-technical book on the subject
is:
"The Art of Genes: How Organisms Make Themselves"
by Enrico Coen
The fact that he wants to invoke fractals in
introns to explain structure suggests that
he dosen't know the molecular biology very
well.
If you're really interested in this sort of thing, you might want to check out something called OOOP , which is a intriguing combination of biology and OOP.
mhack
Building a better ribosome since 1997
On the other hand, if a scientist doesn't patent an idea, a corporation surely will.
Not necessarilly, and even if it were true, if the scientist publishes first, that is prior art and the corporate patent won't hold up in court. Indeed, if the USPTO were not being criminally negligent in its fudiciary duties under the constitution, it wouldn't even grant the patent in such a case.
The Future of Human Evolution: Autonomy
Saying DNA sequences is a 'fractal' (really self-similar) is nice, but not that profound, yet.
In fact if it were a uniform fractal then it would have VERY LITTLE evolutionary mechanistic importance. Only if the law were sufficiently different that physical mechanisms translating DNA patterns into phenotypes (expression of organisms) could account for observed differences in organisms would something behind the fractal law matter.
Indeed, most of the thinking on the introns is involved in eludicating the mechanisms of the "fractal pattern" but this is all in the mechanisms of replication and crossover. There is little science showing mechanistic (and not just correlations) translations back.
As far as patenting "all methods of observing fractal patterns" I point out this prior art:
R Roman-Roldan, P Bernaola-Galvan, J. L. Olivier, "Sequence compositional complexity of DNA through an entropic segmentation method." Physical Review Letters, V80, p1344 (1998).
Here is its introductory paragraph:
"The analysis of sequence correlation structure, in both the spatial and the frequency domains, resulted in the finding of short range[1] and long-range[2] correlations in nucleotide sequences, thus uncovering a complex fractal structure of DNA."
Both[1] and [2] refer to a large number of references each.
The paper then goes on to discuss an information theory based statistical approach to derive an automated algorithm for hierarchical partition of DNA sequences in to succesively more homogeneous regions. The obsevations is that the breakups are self-similar and thus roughly fractal.
Not only that but the trend towards greater 'complexity' with organism complexity is also observed here.
Note that traditionally fractals are defined on a continuous geometrical space, and since there is a minimum breakpoint size (one base pair) in sequence analysis this is not a true fractal---there will only be self-similarity in some "scaling range".
Note I am a physicist in nonlinear dynamics and happen to know a little bit about information theory. I am not a biology or genetics expert.
See for instance the 8 year old research in:
CA Chatzidimitriou-Dreismann, RMF Streffer, D Larhammar (1994), "Are there any fractals in DNA of living organisms"
RF Voss (1994), "Long-range fractal correlations in DNA introns and exons", Fractals, 2(1):1-6.
Professional Wild-Eyed Visionary