Why don’t identical twins have identical fingerprints? | Hacker News Reader

Why don’t identical twins have identical fingerprints?

qi.epfl.ch

77 pointsfgeorgy33 comments

The article doesn't answer the question in any satisfying way. A layman would probably believe that the environment in the uterus is the same for both twins, too. The "satisfying" answer is still quite boring: as with everything else, there are stochastic processes in play during embryonic development. For one, cytoplasm is divided asymmetrically, so the two embryos don't begin their separate journeys with a perfectly equal share of their progenitor's resources.

Also, this is a rather silly sentence:

Although very similar, the fingerprints of identical twins are nevertheless unique because they are not completely determined by genetics: they are an example of phenotype, which refers to the physical characteristics of an individual that are determined by the interaction of genes and the environment.

By this reasoning, the only thing completely determined by genetics is genes, since every manifestation of genotype could be called a phenotype.

> By this reasoning, the only thing completely determined by genetics is genes, since every manifestation of genotype could be called a phenotype.

But this is absolutely correct, so the fact that the sentence leads to this correct conclusion doesn't make it silly. Genotype is just genes. The manifestations are all determined by the interaction of genes with the environment and the manifestations of observable traits together make up the phenotype.

Some aspects of this are practically tightly controlled by genes because the environments which allow an organism to develop at all tend to also be ones in which the genes involved are expressed in certain ways, and some are less tightly controlled by genes, because there is more practical variation in interaction with the environment. But these are differences of degree.

There's no substantive point of difference between my post and yours. I'm simply suggesting that perhaps the "big reveal" for this question is a bit trivial, even for the general public. Everyone knows that no matter a boy's genome, he isn't going to grow as tall as Yao Ming if he only gets 1200 calories per day in his diet.

That is the definition of "phenotype" as I understand it. Phenotype is the observed traits of an organism, while genotype is the genetic (discrete) code in the organism's DNA. Genotype obviously largely affects phenotype, but it doesn't entirely determine it.

>> By this reasoning, the only thing completely determined by genetics is genes, since every manifestation of genotype could be called a phenotype.

I would say half half. Obviously a lot comes from genes, but then environment takes its toll as well.

Robert Sapolsky makes an interesting point about the idea of "a" gene-environment interaction[0]: genes only exist in the context of an environment, and so "The problem with "a gene-environment interaction" is the same as asking what height has to do with the area of a rectangle, and being told that in this particular case, there is a height/length interaction."

His course on Human Behavioral Biology [1] is perhaps the single best science course I've ever encountered. If someone could only take one science course, I'd recommend this one.

[0] https://edge.org/response-detail/25536 [1] https://www.youtube.com/playlist?list=PLD7E21BF91F3F9683

I still love how people simply assume fingerprints are unique even though there has never actually been any studies with a large enough sample size to make that determination

Fingerprinting, at least for the purposes of criminal investigations is pseudo-science at best, akin to the now debunked microscopic hair analysis.

DNA is always given as a probabilistic statement "there is a 1:1,000,000 chance this DNA belongs to someone else"

However with fingerprints, like microscopic hair analysis, it was simply match/no match... a binary state that is given far far far more weight than it should. The perceptions of the examiner weigh heavily on the outcome.

Finger print examining is an art form, not a scientific discipline

How could they be anything other than unique? A fingerprint is a 'function' of a unique individual. You might as well claim that human faces aren't unique.

Bear in mind, of course: the degree of uniqueness between two samples is determined by the metrics used to distinguish between them. I don't know the state-of-the-art with fingerprint analysis, but with a purely physical measure, how can fingerprints not be unique?

At that kind of physical level every time the same individual is fingerprinted you get a slightly different result.

I am saying there is no study to prove they are, thus one must assume they are not.

Further there is not be a study to see if my left index may match your right middle finger, etc.

I do not believe there is enough possible entropy for ~74 billion 100% unique never to be repeated patterns on finger prints... (7.4billion humans x 10 fingers per human)

Many people have problem telling identical twins apart, so under that metric it's true that human faces aren't globally unique. There's even a project to find "twin strangers", which has had some success (see the images at the bottom of https://twinstrangers.net/ , though they also had the same makeup applied, which makes them look even closer).

> how can fingerprints not be unique?

How can the hash of two different strings be the same? It's all part of the Birthday Paradox, combined with that it means two identify two physical objects as the same thing.

Now, at some point no two human-scale objects are the same. Even with two mass-produced objects, there are slight variations in production because of non-zero tolerances. For example, two sequential $1 bills differ of course by the serial number, but also contain minute differences in the paper.

So as you say, we need to distinguish between how fingerprint analysis is done, vs. the most detailed possible analysis. And we know of several cases where fingerprint identification has been incorrect. One is Brandon Mayfield, a US attorney who the FBI believe, on the basis of a fingerprint match, to have been one of the Madrid bombers.

As a result of that case, and similar ones, people have started to ask the police and others for data which shows that "all fingerprints are unique" is really true. This myth started over 100 years ago, but was based on a small sample size. There's never been the analysis needed to figure out the actual error rate, especially if there are, say, 100 million entries in the system.

See also https://en.wikipedia.org/wiki/Fingerprint#Validity . And notice how that Wikipedia page also includes "No two fingerprints have ever been found identical in many billions of human and automated computer comparisons.[citation needed]". Where is that citation?

Probability is not a guarantee.

Randomness does not guarantee uniqueness.

It's not just the fingerprints. Entire organs may grow according to the elegant mathematics of reaction diffusion systems. Not to mention metal alloys, igneous rocks, suns, galaxies, and so on...

A mathematical theory proposed by Alan Turing in 1952 can explain the formation of fingers

http://phys.org/news/2014-07-mathematical-theory-alan-turing...

Plants can be cloned easily, but the leaf and stem patterns don't match exactly. Those are generated by a statistical process and can be simulated with an L-system.

The human genome is only about 3.2GB. All the fine detail can't be in there.

[1] https://en.wikipedia.org/wiki/L-system

> What purpose do fingerprints serve? Apart from improving our grip on objects...

This has been partially debunked[1][2]: The area of skin in contact...was always 33% less than if the fingerpads were completely smooth. This confirmed that fingerprints do not improve our grip, because they actually reduce our skin's contact with the objects that we hold.

I say "partially debunked" because it might improve grip on wet surfaces: They may allow water trapped between our finger pads and the surface to drain away and improve surface contact in wet conditions.

It seems similar to the situation with tire treads. If you ask most people why tires have ridges and valleys, they'll say that it improves grip on the road. But if the road is dry, you'll get bigger contact area and therefore better grip with completely smooth tires. In fact race cars use completely smooth tires for better grip (when the racetrack is dry). Tire treads exist to conduct away water when the road is wet, and they improve grip only in that situation.

In an interview with the co-author of the friction study, he made some interesting speculation about the purpose of fingerprints: My preferred theory is that they allow the skin to deform and thus stop blistering. That is why we get blisters on the smooth parts of our hands and feet and not the ridged areas: our fingerpads, palms and soles.[3]

[1] http://newsvote.bbc.co.uk/mpapps/pagetools/print/news.bbc.co...

[2] http://jeb.biologists.org/content/212/13/2016 (actual paper)

[3] http://www.manchester.ac.uk/discover/news/article/?id=4715

Well, they should test actual grip, instead of percentage of skin contact. So what if it is 33% less? What if it somehow grips better having the prints (as might be the case on wet surfaces)? Also, don't go testing in some material that was never touched upon by early humans.

> they should test actual grip

Well, the referenced article says that that's what they did (in addition to further measurements to determine area of contact).

Upvoted for asking a question in the headline and then immediately answering it in the first sentence of the article.

EPFL in the top 15 unis in the world almost consistently in any of the sciences. For CS and mathematics, they'll usually rank top 5 with INRIA and Cambridge, perhaps higher at the graduate level. (I was going to go to attend for a graduate degree in Algebraic Topology there or but the opportunity cost and, well, I suppose primarily coming to terms with my own mediocrity made me decide against.) But yeah, they're not really trolling for hits to generate ad revenue ;).

Re: one sentence answers - I'm with you. I took this lovely Python module and pipe most everything that I know [1]will be filler to this https://github.com/miso-belica/sumy. I've been doing it for a while now and I don't feel as though I am missing out on anything.

I don't know if this is as significant as fingerprints or not, but I have a coworker with an identical twin, and her sister has Crohn's disease, and my coworker does not. Chron's is apparently genetic, but there must be some environmental factors as well.

or it could reflect a damaged microbiome.

According the Wikipedia: '...dynamical systems that are highly sensitive to initial conditions...'

https://en.wikipedia.org/wiki/Chaos_theory

You might as well ask why two cloned trees don't have the branches and leaves at the exact same positions...

Except that in your example, the trees are influencing by environmental factors throughout their life, whereas the first sentence of this article says it's environmental factors in utero that influences fingerprints.

Fingerprints form while in the uterus. The location of a tree's first leaf will be influenced by the factors up until that time. It's still the same thing. Leaves won't move around from branch to branch.

It isn't a perfect analogy, but it still works. The only reason the other option, "development is completely random" isn't the right answer is because of the word "completely."

Honestly, the survey question is poorly asked, and doesn't produce a particularly interesting discussion. Maybe something was lost in translation.

Which is also an interesting question.

That's not quite as strong of a question, since trees (as far as I know) don't regrow the same pattern of branches and leaves repetitively, yet fingerprints remain roughly the same throughout one's life despite skin constantly shedding and regrowing.

Because not every factor is genetic. Most of them are environmental and random.

The best example is with trees. No two trees have the same structure of branches given that they have the same DNA.

The world is mostly stochastic.)

Why don't laplacian systems with slightly different initial conditions evolve in the exact same manner?

Robustness is drive by homeostatic feedback loops and/or equilibrium states. Some systems will act chaotically and will be more sensitive and unpredictable to changes in initial conditions.

Which gives an answer to the question posed by the post.

I'd like to see how much bilateral variation there is compared to phenotypic variation between twins.

Reply on news.ycombinator.com