Be a Superhero with TeamCFC. Help us perform miracles.

[IglooDame]

Are you saying that there are no paritiy bits in the genome?
No wonder we cease to function after a few years!

So the lack of parity bits leads to failures of the naughty bits? Who'da thunk it?

 

[ironduck]

Oh I've been called worse than fe-quacker

 

[Till]

Without parity bits there's no error proof reading/copying of the other bits. No wonder mutations happen all the time! I hope they plan to fix this issue in humansV2.

 

[Sophie 378]

@Till - No, actually, sorry if I explained it badly. On my understanding of parity, there're plently of opportunities for copy errors to not result in a changed outcome - here's a link to a post about how genes result in proteins from here. If you look at the table of codons for amino acids, there are several nearly-identical codons that give you the same amino acid. Also, the copying gets checked against the complementary strand - overall the rate of copy errors is very small.

Of course, I may have misunderstood you entirely ....

@BCLG100 - I just chose a random sequence of 1's and 0's to make my example of 10 10 01 00 10 00 00 11. DNA isn't eternal; it's got to be repaired if it gets damaged, and replaced sometimes; also when cells reproduce, you need a copy of the DNA for the new cells: so it gets copied.
Copy errors sometimes happen if the cellular machinery for copying DNA has a blip, or if the DNA is damaged (by UV light, chemicals, whatever). A copy error usually means that you get a near-identical strand of DNA, but with errors: in the 10 10 01 00 10 00 00 11 example, I coloured one red to show it was a copy error - a 1 accidentally written as a 0.
The original strand: 10 10 01 00 10 00 00 11
Then makes this: 10 00 01 00 10 00 00 11
Just the one change. Note how everything else stays the same. That sort of error is fairly easy to make, and it's just a little error.
However, sometimes the copying machinery has a big snarl-up, and instead of copying properly, it misses out a base - or accidentally adds one in. This is very rare, but it can happen, and when it does, it gives you a much bigger error.
Using the same original strand:
10 00 01 00 10 00 00 11
It's supposed to copy it to make another
10 00 01 00 10 00 00 11
But if another one accidentally gets thrown in (shown in blue, then everything after that becomes an error.
So
...........10 00 01 00 10 00 00 11
copy as 10 1 then the machinery goes on from there:
.......... 10 100 01 00 10 00 00 11
But these are being read in twos, so you've got to break them up differently:
.......... 10 11 00 10 01 00 00 01 1

Maybe this was a bad example - these were originally bits in a floppy disk, not letters in a DNA string.

Edit in progress because I can't find the darn link I'm looking for.
Done.

 

[Till]

Sophie: I was just joking with my question!
But for the sake of better understanding between chemists and computer scientists here is an explanation of what i meant with parity bits:
They are one of the simplest forms of error checking. You basically just add one bit at the end of your data. The value of the bit is determined by the number of 1's in your data. If you have an even number of 1s, then the parity bit is 1.
Example:
Your data:
00 01
With parity bit:
00 01 0
I'd love to see how mother nature handles error checking, but your link doesn't show up.

 

[ironduck]

So can we agree that two bits won't be enough for the four bases because it doesn't leave a good way to mark for new sequences, etc?

If that's the case, but assuming the markers hardly take up any space, is that the way you get it back again (assuming 6 of the bits in a byte will then be mostly unused) when you pack? If we use 8-bit bytes and we in effect only use two bits per byte most of the time in describing our DNA strand, then we can pack it about 75%? Or is that all wrong?

 

[Sophie 378]

Ah, sorry, showing my misunderstanding of parity there.
My link works for me ... try again http://forums.civfanatics.com/showthread.php?p=3695662#post3695662
Thread 158912 (Do you beileve in evolution? Why or why not?) page two post 37 in the spoiler. Ah hell, I'll just quote it.

Spoiler nucleotides, DNA, amino acids and proteins :

A nucleotide is like this:

The nitrogenous base (green) is variable - the others are constant. IE, the sugar (pink) is always ribose for RNA, and always deoxyribose for DNA. A phosphate (purple) is just a phosphate group.
A nucleoside is the nitrogenous base (green) + sugar, a nucleotide is base + sugar + phosphate, and a base is the nitrogen-containing single or double ring that points into the DNA helix, pairing with the one on the opposite strand.

There are four kinds of bases in DNA:

Note how they are labelled as Adenine, Cytosine, Guanine and Thymine.

(RNA uses uracil (sorry, can't seem to put in as a pic) instead of thymine, and since the sugars are riboses instead of deoxyriboses, the "bases" are called ribonucleotides and are adenosine 5'-triphosphate, guanosine 5'-triphosphate, cytidine 5'-triphoshate and uridine 5'-triphosphate ... but normally, in an RNA molecule, just referred to as adenine, guanine etc - or just AUCG.)

DNA is a very long string of these bases, which pair A-T and C-G like so:

The two strands of DNA are antiparallel (ie one runs up and the other down), and coil into a helix.

The sequence of nucleobases is the genetic code that you hear about. The sequence is transcribed onto RNA, making a reverse, complementary strand: the DNA double-helix opens up so that free-floating ribonucleotides can come in and form a complementary strand. The RNA strand is then translated by ribosomes into a sequence of amino acids to make a protein. Since there are only four bases, and you want a variety of proteins, the bases are read in threes - triplets - with a set amino acid for each individual codon of three.

This figure explains how the RNA is translated: the RNA codon UCA would give you Ser. The ribosomes just throw in amino acids in the matching configuration, sticking them together to form a protein chain. The codon AUG doubles for a START signal.

Amino acids are what make up proteins; proteins are the main structural and functional molecules in biology. They have this basic structure:

The R group is variable again: there are 20 commonly occuring natural amino acids.

Proteins are formed by the condensation of many amino acids - they bond together in a long chain.

A polypeptide is just a long chain of peptides - ie, a protein.

For more info on protein structure, read the Folding@Home links in my sig!

Mutations arise when the DNA is damaged, incorrectly copied or incorrectly repaired. In non-coding regions (junk DNA etc), this may have no effect - but in an actual gene, or a regulatory region, it may well change things. If one base is substituted for another base, you may get just a change of one amino acids in one protein; or there may be no effect. If a base is added in, or removed, then it will snarl up a whole sequence of amino acids, giving a different protein. This protein may be non-functional, or have its function affected positively or negatively - or it may do something else intirely.

 

[Sophie 378]

So can we agree that two bits won't be enough for the four bases because it doesn't leave a good way to mark for new sequences, etc?

If that's the case, but assuming the markers hardly take up any space, is that the way you get it back again (assuming 6 of the bits in a byte will then be mostly unused) when you pack? If we use 8-bit bytes and we in effect only use two bits per byte most of the time in describing our DNA strand, then we can pack it about 75%? Or is that all wrong?

DNA gets around this cleverly by doubling-up on some signals. Eg, the codon AUG means "Methionine", or it can mean "Start" - an amino acid in a protein, or a new sequence signal. There's also a Stop signal I can't remember off the top of my head, but it's in the spoiler in the post above.
EDIT oh darn, the image I want has died! I'll just find another one ...

Fixed

I'm not doing very well here today, am I?

 

[ironduck]

DNA gets around this cleverly by doubling-up on some signals. Eg, the codon AUG means "Methionine", or it can mean "Start" - an amino acid in a protein, or a new sequence signal. There's also a Stop signal I can't remember off the top of my head, but it's in the spoiler in the post above.

I'm not sure what you mean by getting around it.. with digital bits and bytes you need a set number of bits to work as a 'container' of information, right? So if we decide to make a byte 8 bits then we have a container that allows for 256 combinations. But if we aren't actually using most of the 'space' in the container and only use, say, 6 combinations, then it should be able to compress the empty space into basically nothing..

The reason this is relevant (unless I'm misunderstanding which I might well do) is that if we're only using 2 bits per base, but still using a regular 8-bit file system then aren't 6 of the bits wasted if they're left unpacked? Or can you just code several bases into one byte?

As for your example, what determines if AUG means one thing or the other? Is that simply a function of its position or is there a marker that determines it?

edit - ah I see you posted image now, I guess that explains

 

[El_Machinae]

So in effect we have 2 GB of storage in our DNA not just 1.

Kinda. In addition, bases can be methylated (which is one way we can tell which DNA strand is the original and which is the copy). While the DNA throughout an organism will be the same (mostly), the methylation of the DNA will be markedly different depending on which type of cell you have.

Methylation basically is the RAM aspect of genetic computing, indicating which DNA applies in a certain cell.

edit: Sophie, I've always thought that if you can teach something, it means that you know it. So, I can assume this will help you in school.

 

[Sophie 378]

As for your example, what determines if AUG means one thing or the other? Is that simply a function of its position or is there a marker that determines it?

That's one of the things they always say "What a good question" and then recommend a horribly advanced book and/or some papers that will be coming out soon. IE, I don't know, and I'm not sure if it is known. It's "a subject of active research" or something.
I think it may be to do with how the DNA is modified. Real DNA in a cell has phosphates and proteins stuck all over it, making some bits more active, moderating others, etc. Google regulatory proteins! That wouldn't be encoded so easily into a floppy disk!
[edit] Cross-post with El_Mac. Yeah, what he said.

 

[El_Machinae]

As for your example, what determines if AUG means one thing or the other? Is that simply a function of its position or is there a marker that determines it?

It means both. If the methionine was essential to the protein, the protein is sometimes made. If it was supposed to be a stop codon, when the stop is ignored then a glob of useless material is made (and degraded).

In addition, the codons before the AUG often matter, because they'll coil the RNA in such a way as to prevent (or encourage) stoppage. Finally, the DNA can be altered (physically) to encourage (or prevent) stoppage.

"In the beginning" it really looked like a fuzzy system. But, as time passes, the 'fuzziness' is fading away (but it still seems fuzzy)

 

[ironduck]

edit - double post

 

[ironduck]

It means both. If the methionine was essential to the protein, the protein is sometimes made. If it was supposed to be a stop codon, then a glob of useless material is made (and degraded).

But how does the process 'know' whether methionine is essential to the protein or not?

 

[El_Machinae]

Well, sometimes there is backcoding or DNA alteration (that changes the AUG into a Methionine codon).

However, a lot of the time it seems like fuzzy luck. Sometimes the process will read it as a STOP, sometimes it will read it as a MET.

If the process is wrong, the errant product is degraded eventually. If the process is correct, the protein forms normally.

Basically, it looks like that enough are 'right' such that the cell can continue to survive.

 

[ironduck]

just fiddling about the interweb, maybe someone else will enjoy this http://nobelprize.org/medicine/educational/dna/index.html

 

[Till]

Ah, sorry, showing my misunderstanding of parity there.
My link works for me ... try again http://forums.civfanatics.com/showthread.php?p=3695662#post3695662
Thread 158912 (Do you beileve in evolution? Why or why not?) page two post 37 in the spoiler. Ah hell, I'll just quote it.

Thanks Sophie! That is very informative. It looks like there is a fair bit of redundancy in the table, which puts my mind at ease some. The system behind me escapes me at the moment, but it is late night already, after all.
For instance, you can stop reading after you read "GG", you already know that you'll get "Gly". But if the first "G" is mistranscribed as an "A", you suddenly get "Ser". That's not too failsafe.

 

[El_Machinae]

Yeah, but you'll notice that anything after the GG doesn't matter.

In addition, Ser might be similar enough to Gly, that the protein will serve the same function. This concept is called homology, and is one of the reasons why we assume that species are related when speaking of evolution.

 

[Sophie 378]

Another fun aspect is "wobble" - the RNA translating machinery that makes the proteins reads the first two bases of a codon more carefully anyway.

 

[El_Machinae]

I'm remembering now, why my genetics courses seemed to exponentially get harder ...