Slashdot Mirror
Portrait Sculptures From Genetic Material
rogue-girl writes "Artist Heather Dewey-Hagborg showcases portrait sculptures from genetic material collected in public spaces. DNA extraction and processing are done in a DIYbio-compliant fashion at the DIYbio hackerspace Genspace in Brooklyn, the collected information is then given as input to a 3D printer. The software developed and used for this project is awkwardly dubbed 'friendware', that is it is neither open nor closed, but only available to friends. Reconstructing faces from DNA is not new: scientists already successfully reconstructed Neanderthal man's face from ancient DNA back in 2008. At first sight, the artist's project may seem fun and quite impressive as high-voltage science proves once more feasible at home, but all the data one can have access to from totally banal samples leaves open worrying perspectives about how easy it is to use DNA collected in public spaces for "fingerprinting" people against their will and without their consent."
That's all I've got to say.
It is unwise to ascribe motive
I have had contact with Manu Sporny, and his portrait doesn't look like him. You can google him to see for yourself. Of course, I imagine this kind of tech will only get better...
Look no further then Snooki!
The software developed and used for this project is awkwardly dubbed 'friendware', that is it is neither open nor closed, but only available to friends.
So.. closed source, then.
An enigma, wrapped in a riddle, shrouded in bacon and cheese
I notice that they took these samples from cigarettes and chewing gum. Seems to me that if you leave something like that in a public space, there's no privacy concern.
Lifting off of a used glass/hair follicle/sweaty towel at the gym/etc. would be a bit more worrying.
But since a person's features are more than their base structure, it's probably not too big an issue anyway. It's highly unlikely that they'll be able to model exactly what a person looks like at their current age/health.
This method could definitely help with missing persons issues though, as an adult model could be created based on a child's DNA that would look "similar" to the actual person.
I'd be interested if two runs on DNA samples from the same person would turn out faces that look the same....
Reconstructing faces from DNA would be new. Nothing of the sort has ever been done. And how could it? It's not only about genes, but also to what degree they are expressed.
How many of you immediately thought of that scene from "There's Something About Mary"?
I wasn't aware dirty hairs and face was a genetic trait. I guess then the Neanderthal became extincted as soon as the soap was discovered.
Achille Talon
Hop!
Thanks a lot. I may never get that image out of my brain no matter how many times I wash it.
Have gnu, will travel.
And this shit won't happen!
I don't believe this so called "artist", and think that she is, in reality, nothing more than a self-promoting B.S. artist.
DNA testing isn't cheap under the best of conditions, when you can take a blood sample in an office, and in the case of recovering genetic traces from discarded cigarette butts, etc. that are full of environmental contamination, the obstacles to acquiring a usable sample are formidable. I would bet that this "artist" cannot provide any laboratory reports or paid receipts from an accredited DNA laboratory to back up her outlandish claims of having extracted and analyzed these supposed genetic samples.
Moreover, her claim to have written custom software that can reconstruct facial features, and actually synthesize a 3D face from the results of testing two or three genetic characteristics is laughable on its face.
I can only assume that she has personal connections with the media outlets that are promoting her fake art, or she has paid a PR firm to plant these ridiculous stories in willing press outlets.
on my girlfriend's lower back. She was not impressed.
Obvious practical joke. Not even close to being believable.
Knowledge is power; knowledge shared is power lost.
No Open... to the highest bidder. Then closed.
Sure, your genes have a lot to do with how you look. That's why identical twins look so similar. However, we are absurdly far from being able deduce facial features from genes. Even in the linked article, most of the information in the Neandrathal reconstruction came from discovered bone shapes. What did we learn from the genes? Maybe something about skin pigment and possibly hair color. That's it. Nothing about the facial structure came from the genetic information. So any fantasies about pouring genetic material into a box and then finding a corresponding picture of a face on a monitor... that will probably be a fantasy forever.
Imagine, several years from now when the technology has matured, a woman in the police force whose team has gathered several DNA samples from a crime scene.
The team analyses the DNA and produces 3D facial models of each suspect, at different ages. Then they use another software to cross-reference key facial points in the 3D models with the already-existing photo database of every US citizen, along with their personal information.
Bingo, the policewoman just got a nice list of suspects to interrogate for the crime, using existing resources, and without the politically-difficult need to build a DNA database of every citizen.
Other than single-gene diseases, we have a very poor idea of which QTLs influence which traits (assuming QTLs even exist). With billions of possibilities among all the alleles and haplotypes, one can make up whatever combination of DNA letters they want and "prove" its association with a given phenotype through a low p-value. One can also use "penetrance" to explain away deviations. One can build whatever Bayesian models they want, building a house of cards that will collapse very quickly as new data is introduced which contradicts the old. I was going to do my dissertation on statistical genetics, but the more I learned about it, the more I learned how intractable the problem is. I did my dissertation on a different topic.
It would be nice if they could compare the reconstructed faces with real faces.
So any fantasies about pouring genetic material into a box and then finding a corresponding picture of a face on a monitor... that will probably be a fantasy forever.
Actually we can do that now. It just has a rendering time of about nine months.