Biotechnology Corner

In the truly random string of DNA, there is a random chance that there will be patterns.

In the worm DNA, there may be patterns, but then again, there may not.

HTH.

HTH?

I suppose then if you were to look for the statistical differential between a lowest-entropy 'random' strand you're no better off than you would be just looking at the worm DNA on it's own.

That's kind of what I was getting at. Will there be a higher level of organisation, patterning or structure in the worm DNA compared to the random DNA, regardless of whether a random strand has some patterning of it's own?
(Which, with anything that large, there will be patterns - look at the Bible Code)

Extending the Bible Code analogy - would there be more patterning in an 3D matrix of information which has been written and is intelligible, or in a matrix of identical dimensions, but made of randomly ordered letters?

Doing instinctual big-picture statistics, you'd think that the two would be comparable...but with the DNA...would they be?

What you'd be looking for would be the deviation from the expected random. There are statistical tools for denoting whether a magnitude of deviation from an expected result is significant or not. I was just wondering if there might be a way of manipulating those tools for this purpose.

what exactly do you mean by patterning? all you really need to form proteins is a start (AUG methionine) and stop amino acid. so if you have those in the random, you could conceivably get protein formation. three codons code for an amino acid and there is also something called a wobble that allows for the an amino acid to be formed for without exactly the proper code being present.

there is a huge amount of randomness in any genome. Think introns and exons. not everything codes for something. and actually there are a lot more non-coding regions than coding regions in any given genome.

am i making sense or just babbling?

By patterning I mean a deviation from the random - structure, order, organisation.

I think you're making perfect sense, but what I was aiming at was not the identification of coding regions for the generation of proteins, nor the generation of proteins through the actions of amino acids.

As far as I understand it there is an enormous amount of interaction between different parts of the genome which is not understood, that straightforward coding is rare anyway. Identification of Exons and Introns and the various interactions of mRNA, to me, would provide an interesting extra layer of information to any discernable ordering in the DNA, however scance and widely distributed.

Now, the real test...am I just babbling?

Summed up, what I wanted to say was: what would be the differences in structure between a random strand of faux-DNA and an evolved, functional one?
and
Could we tell anything from this?

I can't believe I actually read all this.

Statistically random would mean you would have equal numbers of A,T,G,C in the polynucleotide chain. How does that compare to DNA from living cells? Eh, not much difference. Does that mean they are equally entropic? (Is that even a word?)

In a complex organism, the same DNA is in every cell, but the cells have widely divergent functions. The same section of DNA could be an intron or an exon, depending on the cell (and the proteins that fold it into weird shapes, exposing what needs to be expressed, and hiding what doesn't). The process of phenotypic expression is rather well-understood.

This is all a long-winded way of saying, you aren't asking a very clear question.

(I used Molecular Biology by David Freifelder when I was in college studying Microbiology. Excellent book for this sort of thing.)

Gotcha, Gopi.

It was the process of phenotypic expression which I was getting at. Not necessarily the mechanics, which I think is well-understood*, but the motor for it.

DNA generates different proteins and hence expresses the vast array of different cellular types inherent in the phenotype as a whole.
Check.
I was curious to see if the structure of DNA shows coherence enough to be differentiated from randomly-sequenced G,A,T and C polynucleotide chain in a practical sense.

For example, this structure of mind-boggling complexity is, nonetheless, identified as more ordered than random, and has a lot in common with another structure further down the chain, which indicates that there might be a relationship here of which we weren't previously aware.
Hypothetically.
This is the sort of outcome I was wondering about.
Hypothetically.
We're talking epic statistics here, which I wasn't even sure could be done, or if they were practical.

I didn't even know about introns and exons until Jillian left her comment, but now I see I've asked, in a circuitous way, a question to which the answer is already known.
The possibility that there might be a significant statistical difference from the random- clustering,repetition, or other patterning - (and that this may in turn tell of the complex relations between different stretches of DNA) is by far secondary to the actions of identified introns and extrons (depending, as you say, on the cell type).
My question was a roundabout way of asking 'How could we find out which bits of the genome relate and interact with other bits?' when a question further down the line, namely 'Why do these bits do what they do?' was already out there.

Thanks everyone. I have more reading to do.

* Although not, I should hasten to add, by me.

I think it's your fan belt meself...

Re: interactions among alleles, how do you find out? Single-variable analysis. Quite simple really. Then you get into complexities like co-dominance, pleiotropy, and multi-genic phenotypes, but the idea is still the same. Isolate and test.

Thanks for forcing me to remember this stuff. I haven't even thought it in years. Since college, really, since I don't work in this field any more.

Thanks to you too, Gopi!
I'm just really getting into the subject. Jillian's given me some good links via email...

To go back to the original way you asked your question:
To compare the randomness of the nucleotide sequence to the actual sequence observed, Biology students are taught to consider that at each nucleotide position in the sequence there is a 1/4 chance that a given nucleotide will be at that location. Then say, for a sequence of 4 nucleotides there is 1/4 x 1/4 x 1/4 x 1/4 = 1/16 chance that the sequence would occur by chance. For a longer sequence the number would be 1 in something greater, so I suppose this formula could make some sense, but doesn't there seem something not right about this? It just doesn't seem logical to me.

Also: There is no need to correct yourself for the question you asked. People might already be asking’ Why do these bits do what they do? but asking 'How could we find out which bits of the genome relate and interact with other bits?' first would be a good idea. Gopi seems to think that things to do with DNA expression are well understood, but are they really? If he (and others of this opinion) thought about it on the level at which you raised your question, it would become apparent that actually the field is not so well understood.
Gopi- I would be glad if you can prove this not to be the case.

Isn't 90% of our DNA junk anway? (Yes I know, mine is probably more junk than most ha ha can we move on now)

I'm not sure which is worse... That someone asked a question like this in a blog, that it actually resulted in this many comments, or that I found the entire thread so damn interesting.

42

LJM; I think the 'something is not quite logical about this' feeling might originate from the point that Pete made at the start - the odds of one particular configuration of a 4-nucleotide string is 1/16...but there is a 1/16 chance for every other permutation as well.

If the nucletotides are arranged randomly.

The thing I was getting at was wondering if you could spot arrangements of nucleotides which bucked the statistical trend so much that you could mathematically prove that they were not arranged randomly* (which, after several million years of evolution, we know they weren't). From that I wondered if you could then cross-reference areas of similar 'non-randomness' to see if there were relationships between them.

Maybe that knocks out that sneaky illogical feeling?

*Still bearing in mind that there would be an amount of apparent order in any set of random data, let alone in the enormity of a randomly-generated strand of faux-DNA

Oh, and Gopi, the whole cross-reference thing was an attempt to try and map some of the complexities you mention...without single-variable analysis...or at least to use as a guide as to where to direct it.

Adrian, yeah, there's a high percentage of 'junk' DNA, not as high as 90% though, I think (anyone?). As Olly points out, those genes could be thought of as hitchhikers.
.
.
.
.
.
.
.
.
.
.
.
*cough*

Hmm, so it would appear that really bad, stupid attempts at jokes can kill scientific conversations.

Interesting.

I thought you were apologising for having such deep thoughts that make my Monday brain short circuit.

I have no clue where to point you but tell me what's on the summer burn CD anyway.

There was a joke in there? Seriously?

Not conforming to any of the rules of comedy, no.

You mean like

Rule 1: Be funny.
Rule 2: Makes sense.

Those rules? :-)

Those are they, yep.

Actually, lame as it was, it abides by Cleese's three laws of comedy:

1. No puns
2. No puns
3. No puns

Have you tried kicking the tyres?