Slashdot Mirror
DNA-rainbow, A New Vision of Human Chromosomes
An anonymous reader writes "Two scientists have rendered amazing pictures using datafiles from the human genome project. They assigned different colors to the DNA and rendered images showing interesting patterns and strange structures of our chromosomes. It might be a groundbreaking new idea for displaying and maybe better understanding our genes. With its fascinating pictures it is a beautiful mix of science and art."
No matter how hard I try, I can't see the sailboat!
This is the same principle as the Bible Code which has been shown over and over to be rubbish. If you line things up in various ways you can find just about any pattern you want given sufficiently long input.
Want to improve your Karma? Instead of "Post Anonymously", try the "Post Humously" option.
If you put enough random data together, you're bound to see patterns every once in a while. I bet you didn't know God scribbled pictures in the Bible, too.
I tried banging the side of my computer, but all I see is static. Something must be wrong with the rabbit-ears on my modem.
...are heavily fragmented. This could degrade performance in creating offspring.
Would you like to optimize your chromosomes?
[Yes] [No] [Cancel]
Taste the rainbow!
Only a white page with nothing on it...
:(){
It looks like the DNA has been Slashdotted.
:)
Hopefully the next version will have developed a natural defense mechanism to handle the strain Slashdot puts on servers.
This is about as useful as assigning a different colour to every ASCII character and then viewing the works of Shakespeare, arbitrarily wrapped at 3500 characters. It's not beautiful, it's not insightful, and it's not worth anyone's time.
My genes! They've been slashdotted!
I need tissues!
Actually, does anybody have other good Science/Art websites they can share? I remember having a book, "On the Surface of Things" I think, that basically had lots of colorized/slightly manipulated images from science and technology. Some the shots were magnificient, surprising,and intriguing all at once. I had always thought that sort of thing would be a good tool for educators to get children (or adults) more interested in science. On a side note, I also wanted to set up a website community to bring together artists and scientists to see what how they might collaborate. Never got around to it of course, but has anyone seen anything similar?
I demand royalties
Nothing witty
not much use to have a slashdot story that is password protected.
just my two cents
It doesn't matter what the pattern is, nor what it means. If the pattern is there, then the pattern is there. What does matter is what you DO with the pattern, and maybe why it is there.
Any pattern can be modeled in an algorithm, and from this algorithm it can be extrapolated. A set of data without any patterns is noise; random data. An algorithm found in a dataset speaks of a function, and understanding functions in the human genome leads to better understanding of what we truly are.
All rites reversed 2010
http://www.mirrordot.org/stories/09b848a8bc7c51c52 d483db9d2f1cfba/index.html
Not very spectacular, IMHO.
CC.
TaijiQuan (Huang, 5 loosenings)
I don't think there are any meaningful patterns to be found in a two-dimensional projection of the data. Maybe there can be found something interesting if the data is arranged in three or more dimensions.
I used to think of the DNA as a kind of a programming language for the physical laws that exist in the universe. DNA in its very basic function is a mechanism to assemble complex organic molecules from simpler molecules and / or atoms, so I'm not sure wether we can extract any information from it using a (2d/3d) spatial arrangement like in these images.
Maybe a tree-dimensional approach reveals more details / information about the structure of our genes, maybe we even need four dimensions (i.e. include the flow of time in our considerations).
I dont know what yer talkin about, eh? ARe ye just trollin for the grammr coppers, or di d you just think that yer postage would actally put the fear o' God inter the submiter's braain? Braains. . . Brains . . . .
Always going forward, 'cause we can't find reverse.
Even if it doesn't mean anything, putting false colors onto DNA images (or, for that matter astronomy pics) could have a positive result of attracting interest in the field.
And sometimes, they're just nice to look at!
Piet is an 'esoteric' (useless) programming language that reads bitmaps as source files.
n guage)
... what would the binaries do?
http://en.wikipedia.org/wiki/Piet_(programming_la
http://www.dangermouse.net/esoteric/piet.html
It'd be nice to be able to load the chromasomes up into the piet interpreter, and see what comes out!
Wouldn't it be interesting, though, if it turns out that the genome could be understood as a 'program', and a specially coded interpreter could process it...
I've seen that before... when my TV was on an unused channel and someone started blow-drying their hair.
When our name is on the back of your car, we're behind you all the way!
DNA is very very very difficult to search and index effectively, especially since scientists are very interested in finding sections that don't quite match.
A good friend of mine (hi Paul) has been working on hardware and / or software searching algorithms for a couple of years now. I used to live over his back fence, and he's talked me through a couple of his ideas.
<surprise> Oh, I see he filed a patent. </surprise> Well I can't say any of that was obvious.
09F91102 no, 455FE104 nope, F190A1E8 uh-uh, 7A5F8A09 that's not it, C87294CE no. Ah! 452F6E403CDF10714E41DFAA257D313F.
This will certainly have put the authors' gizzajob plea in front of many eyeballs, and that may be its primary value. A more interesting approach to the harnessing of our pattern recognition abilities to spotting significant sequences in the chromosomes would be to display the genetic code in colours relating to, e.g. the hydrophilic/hydrophobic nature of the encoded amino acids. I agree with earlier posters; anything you spot in an arbitrarily-wrapped 4-colour mapping of bases is so far separated from a meaningful biological message that the site as it stands is just a bit less interesting than zooming in on bits of the Mandelbrot set. FRACTINT, anyone?
It's what the data segment of your app looks like when you accidentally dump it to vga video memory.
boycott slashdot February 10th - 17th check out: altSlashdot.org
Seeing structure where there isn't any Quote :
Strange structures (close your eyes just a little bit to see more details)
C. Sagan : A demon haunted world:
http://www.amazon.com/gp/product/0345409469/
visit randi.org
Who's ever those genes are - he or she is not on the list. Move along.
But nice to see that they have upgraded Prof. Suresh's software to do colours.
Mit der Dummheit kämpfen Götter selbst vergebens
I don't even see the genes anymore - just blonde, brunette, redhead...
In Soviet Rush, today's Tom Sawyer gets high on you.
If they kept the 1D data in 1D and enoded it as music or sound we could use all the technology developed in Digital Signal processing and come up with even more bizzarre stuff. For example the DNA mole could actually sound "Om bhur bhuvasuvaha, Om tats vidhuvareNyum, Bhargo dhiivasya dheemahi, Diyoyona prasaadayaad". Just have to select proper notes and pitches.
sed -e 's/Chuck Norris/Rajnikant/g' joke > fact
If you stare at a graph long enough, you can make it have any pattern you want.
There is, of course, much ongoing research in finding mathematical patterns in DNA. I had a paper published about how DNA SNPs seem to follow a Poisson process in their distribution. Does someone know a good way to visualize Poisson processes? When graphed as they do, it just looks like a sequence of randomly spaced dots.
Overall, I think this is wicked cool, but amateurish from the standpoint of science. Actually, I'd like to see a Gerald Edelelman approach to handling and analyzing the DNA -- which would be wicked cool! See From Brain Dynamics to Consciousness to see what I mean. Applying his neural darwinistic approaches to DNA would not only reveal many surprises, but would be referentially cool, applying neural evolution to what was the result of biological (and memetic) evolution!
OK, so you think I am mad as a hatter. Perhaps. Perhaps not.
Ruby Neural Evolution of Augmenting Topologies
I've been doing this for years with large contigs to help visualize repeats. You'd be amazed at how good we humans are at picking out patterns visually.
It's not really much different to dumping binary data to screen memory. Some old home computers used to use screen memory as a temporary store (e.g. when loading a large programme, prior to relocating it elsewhere in memory), and you sometimes saw interesting patterns in it (ignoring graphics data).
[Happosai]
proof
(and it's also more artistic than linux)
intellectual property law is philosophically incoherent. it is your moral duty to ignore it or sabotage it
...Is this page part of an ARG?
m e
http://en.wikipedia.org/wiki/Alternate_reality_ga
Those patterns look like random data in video memory, with the default color palette of VGA's mode 13h. Ten years ago I wrote some x86 assembly code with quite similar results! :)
`echo $[0x853204FA81]|tr 0-9 ionbsdeaml`@gmail.com
"We share 99% of DNA with the shimp right (now)."
We share a large (not 99%) of our DNA with shrimp (and shimp, whatever those are) because most of it is involved with cellular functioning, you idiot.
Do you use a special keyboard for the completely retarded?
...the difference between geeks, artists and art lovers is clearly illustrated. Those images do not look beatiful to anyone outside of the geek/scientific community. And I'm sure that even within that community there are those who had the same reaction I did. "Hmmm... just looks like the noise filter from Photoshop or the GIMP". Which brings up an interesting question. If we took one of the noise filter outputs and translated it back the other direction, would we wind up with any genomes? ;P
-"...bad old ideas look confusingly fresh when they are packaged as technology" - Jaron Lanier (Digital Maoism on Edge.o
And he is speaking in COBOL
..........FULL STOP.
This publicity-stu^H^H^H^H^H^H^H^H^H^H^H^H^Hexperiment is teh crap because:
FATMOUSE + YOU = FATMOUSE
if they had used Tom Cruise's DNA the resulting rainbow pattern would probably approah a visual singularity.
So, which colors represent superpowers?
- 2 dimensionally
- 3500 points per display line
- left to right
- top to bottom
Sheesh. How about 3 dimensional spiral rendering or spherical or (for the "flat worlders") a cube? Granted, even with the present limitations, I can see some patterns in the data, but then I see Elvis at the Walmart. To paraphrase a famous astronomer : "The humane genome is not only weirder than we imagine, it's weirder than we can imagine".The Russians have won. They have made the world a cesspool of distrust, greed, fear and hate.
If they had made the images different widths, the picture would be completely different. None of the "images" you see in here are beyond anything I'd expect from random chance. Sure it's kind of neat to look at our entire DNA sequence laid out like that, but keep in mind that just because you can kind of make out your dead grandmother's face in one of the pictures only shows that you have a normal human brain.
I made a program in VB6 years ago that can make much the same images. But these are just computer generated by random. Probably would end up being a slug or something. http://www.planetsourcecode.com/vb/scripts/ShowCod e.asp?txtCodeId=32601&lngWId=1
--Edward Dassmesser
Looks more like my early attempts at programming in mode 13h in DOS.
What would happen if you just displayed four random colors? I wonder how many patterns you'd see. Anyhow, I doubt that at the level of the four basic ACGT elements, there is any obvious information
Currently hooked on AMP
Looks like they've finally found the gay genes.
The Rise and Fall of Online Community
Or, you know, you could look at a depiction of the Human Genome that contains real, usable data. Heaven forbid. This is like the folks who make "music" by assigning 4 notes to the 4 different DNA bases and playing them in one long string.
Jerm
Oh, you're not a real doctor, are you?
Actually, he isn't all that -very- wrong - the genome is like a machine code - literally assembly code to generate proteins at slightly higher level, humans at highest level. What we see while watching the genome is about the same as we see while looking at a binary of a program displayed in ASCII. In one hand, it IS all, in the other we can't make out any sense of it - but we DO see patterns. At least in some programs, some of us do. Watch the binary, and you'll see patches of more ASCII printable characters, long rows of NULLs, similar slightly differing short sequences repeating multiple times, forming diagonal patterns on the ASCII display and so on. These relate to copy/paste code patterns, inlined data, initializing data structures with multiple empty strings, zeros and null values and such. I doubt evolution wrote some clean form of for($i=1;$i=5;$i++){hand.appendFinger($i)} and much more likely used dirty cut&paste style of inline addFingerToLimb(hand1, finger1); inline addFingerToLimb(hand1, finger2);inline addFingerToLimb(hand1, finger3)... which can easily create such patterns of repeatablity.
45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
So you have genes of n codons in length, repeated m times, with k interons of length between 1 and l - each codon expressed in base 3^4 - what is the isomorphism to colored light?
However, it might just be their frontpage that's hammered.
You can get 'deeper' pages at http://www.dna-rainbow.org/chromosomes/X.html where X is the chromosome number (1-22) or x or y (lowercase).
-Styopa
So, they gave each base-pair a color? What on earth is the point? 98% of that sequence doesn't do anything. And why is a virtually random sequence of pixels of 4 different colors "beautiful"?
I can understand if they took two different genomes from the same species and did some kind of comparison: different colors for matches, indels, translocations, silent/synonymous/non-synonymous SNPs, etc. Or translated the sequence and colored by hydrophobicity/charge/polarity/whatever. Or showed haplotype conservation between species.
At least that would tell you something, this is just a bunch of pixels with no meaning. A vaguely similar thing I've done was to plot plot SNP density (as color intensity) over the genome - but that was for a specific project, I didn't realize such things are "new visions".
There are definitely prettier visualizations out there too: http://acg.media.mit.edu/people/fry/genomevalence
Even this is a lot more informative (I think www.visualcomplexity.com was mentioned on slashdot a couple of years ago).
sic transit gloria mundi
I saw someone mention "you'd see patterns in any random string" and I saw someone mention a quote from the movie Pi (the quote probly not coincidentially sounds exactly like the up-coming movie The Number 23).
Weird, but I just came across this last night: Play with Pi
[/sarcasm]
I think we all agree there are paterns & structure in the DNA. The point was "half close the eye to see structure".
C. Sagan : A demon haunted world:
http://www.amazon.com/gp/product/0345409469/
visit randi.org
I think he meant chimp, not shrimp.
- None can love freedom heartily, but good men; the rest love not freedom, but license. -- John Milton
- None can love freedom heartily, but good men; the rest love not freedom, but license. -- John Milton
So you think we've sequenced the entire genome? By what definition of "entire?" Ever hear of centromeres and telomeres?
- None can love freedom heartily, but good men; the rest love not freedom, but license. -- John Milton
It's not exactly ground breaking. I mean, i've done things like this when looking at data. The stripes are not 'novel' either. A lot of genes have repeating areas. These encode for repetitions of amino acids that fold into alpha helices and beta sheets, which form large parts of the structures of enzymes. So when you encode the dna as colors, of course you'll see stripes
Collagen genes also encode for the collagen helices, so they will show repitions as well. Same with other structural proteins, such as tubulins, and channels. But it's not earth shattering.
I came across this almost as soon as it was posted. It was already slashdotted, but immediately gave me ideas all the same. I wondered what DNA data would sound like represented as musical notes. Since I have some code modules I developed previously to generate musical notes, I knew that I could have something up and running quickly enough.
The next thing to do then was to find a source of raw DNA data, and that was not so quick. I found where they have this "Gene Boy" teaching toy that explains about DNA transcription. I was also able to steal raw data from it by a simple copy and paste.
Before even trying to learn anything with all the resources available at this site, I quickly wiped up an algorithm to play the data found in the Genome example from Gene Boy. Next was the choice of how to represent the musical notes. Now - I am not a musician by any stretch of the imagination, but at least I know when notes clash or harmonize. I decided that I would experiment with some note assignments and just see how they sound. The proteins are represented by G, A, T, and C so begin I simply assigned G4, A4, B4, and C5 to represent GATC. I set my code up to play fast, and it sounded like a jig. Then I tried playing all the other files available with Gene Boy, including plain, random data. Well, they all sounded much the same, so obviously I needed a better approach. It was time to read the material and learn a bit more about DNA transcription.
I soon learned that it is codon triplets that in the end get transcripted, so its useless to look for patterns at the level of single nucleotides. Only certain regions code anything meaningful, and the rest is basically garbage from the perspective of trying to find audibly meaningful patterns. However, though I quickly learned how to identify the start and end of "Open reading frames" - regions that code actual proteins, I also learned that there are stretches call introns and extrons, and that the introns do not contribute to the final protein, being dropped in the end by the mRNA.
Finally, I have come to the conclusion that it would probably be much more rewarding to use the sequences from the actual proteins produced by the mRNA transcrition process, and that is what I will do next. If anybody wants to collaborate with me - especially - a musician, send me an email: gtaylorATmagmaDOTca. In the end we will post our results. And oh, by the way - I finally saw the images and I think they suck bad. The author of that site, being a biology graduate, could have shown much more insight and imagination. I would not hire these people.
There's a lot more out there, take a look at the repeat analysis at: http://www.4g.soton.ac.uk/~new/genomeReport/ Also a tool called Reputer is commonly used for genomic repeat visualisation..
Maybe if we have them represent base-4 numbers, nibbles of ASCII, or some other numeric base, we'll find a circle, or just a stereo-optic starfish image to cross your eyes over.
Then, this would will be irrefutable proof of something, some sort of README.first, or just random gibberish for monkeys to type out.
Contact.
These are essentially raster scan renderings of a 1-D 2 bit signal. As folks have noted the authors chose 3500 at the raster line size. They also noted that no matter what the raster line size is, patterns appear.
I believe that the source of the patterns are long repeated sequences. A sequence that repeats after a constant number of base-pairs will appear as a diagnonal line. This is independent of the raster line size, however the slope of the line will differ dependent on raster line size. Arcs and circles represent sequences that repeat after a variable number of base-pairs which varies by some polynomial.
I knew it... holy crap... what have we unleashed with our hex editing of the genomes? Hey look I can play online with a blue cat now though! ;) (meanwhile entire system fails) I would like the clover back in my backyard now please... (or at least someone to tell me why I can't find any)
The other post beat me by two minutes. When I saw the article, there were no responses.
Lies! We actually share 100% of DNA with the shrimp right now - they're both made from the same basic four molecules. Oh noes!
Holy cow! I think I see the Virgin Mary in chromosome 15!