Derive time of conception or birth from DNA analysis?

December 12, 2007 11:15 AM

Are there any studies, past or present, devoted to the idea of deriving precise time of either conception or birth from DNA analysis? For conception, I can imagine a science fiction author describing a rhythmic pattern or pulse emitted by something in the DNA whose beats are somehow recorded somewhere, like the veritable second hand of a clock. For birth time I'm a little fuzzier on the possibilities, but one idea that occurred to me was the studying the effects of direct light (sun or artificial) on a baby's chromosomes when exposed for the first time (not counting any rays that might penetrate the womb through the mother). There is also the telomere clock to consider, about which I know very little.
posted by christopherious to science & nature (16 comments total)

Telomere length has been implicated in longevity (not time from conception), but it might be worthwhile to note:

1. Different people start with different telomere lengths
2. After birth, different tissues in a person have different telomere lengths
3. We don't know much about telomere regulation and maintenance

It seems it would be difficult to extrapolate a clean function of length to time-from-conception.
posted by Blazecock Pileon at 11:35 AM on December 12, 2007

No
posted by Megafly at 11:37 AM on December 12, 2007

Plus wouldn't telomere length depend on the number of times a given cell has divided? In which case, wouldn't some cells in a body have different lengths than other cells in the same body?
posted by aramaic at 11:38 AM on December 12, 2007

Interesting. So it would seem that telomere length probably isn't where to look. I had noticed that cloned material and creatures were found to have "older" telomere clocks which immediately cast a shadow on this method for me, although I gather that we still have a ways to go before we understand it well.
posted by christopherious at 11:40 AM on December 12, 2007

Mutational rates can be used as 'molecular clocks', but this only works over long periods of time (millions of years). There's no good way to derive age from DNA alone.
posted by chrisamiller at 11:53 AM on December 12, 2007

Just throwing this out there, but it might be possible to use a radioactive isotope of phosphorous to determine age from DNA, but it would take a lot of work. As DNA uses phosphate groups in its backbone, it would theoretically be possible to introduce radioactive isomers of phosphorous (P32 or P33, with half-lives of 14.3 and 25.4 days, respectively) into this backbone, and measure decay over time.

Of course, you'd have to introduce the mother to a very specific environment where radioactive P isotopes were abundant and then somehow measure the rate of uptake into cells (taking into account that the uptake in a fetus could be significantly different than a full grown adult). Then of course you'd have to ensure that the distribution of P was similar among all fetal cell types (or determine what the distribution is and how to account for it).

Then, after a certain amount of time in a radioactive P-rich environment you could place the mother back in a normal environment and 'start the clock.' By sampling DNA for the level of your radioactive isotopes, you could potentially determine how long it has been.

Overall its somewhat similar to how Carbon dating works, but given the much shorter half-life of Phosphorous it is something that would be measurable in human time frames (C-dating uses C14, with a half life of over 5700 years, making it unsuitable for measuring times withing a human lifespan). Of course, the half life of P is so short that you would only get meaningful results for 1-2 years at most, and probably less than that as after that you'd be down to such minute quantities that it would not be easily measurable.

So potentially possible, sure. Likely to happen, probably not. I'm sure there are other ways that people will think of though, and within our lifetime I wouldn't be surprised to see it happen.
posted by langeNU at 12:17 PM on December 12, 2007

You might be able to look at somatic mutations. There seems to be a fair number of studies that come up when searching "somatic mutation rate" in google and pubmed which might be of interest. This won't be very precise as there are all sorts of (and often not understood) mechanisms involved which cause these mutations and others which can repair them.
posted by Durin's Bane at 12:20 PM on December 12, 2007

Telomeres are implicated in the longevity of DNA, as Blazecock noted. Basically, as a cell divides it makes a copy of its DNA, and portions of the very end of the DNA must be clipped due to the copying machinery; telomeres have no genetic information, and so can be cut off without worry. There is also the enzyme telomerase which is responsible for adding back these clipped segments, because if the telomere ever gets too short, the cell risks actually cutting off real genetic information and becoming mutant-prone, so it is programmed to die.

As you said with "cloned material and creatures," it has been shown that cells sitting in a test tube for a period of time before being implanted and grown into a full organism have on average shorter telomere segments, and thus may live shorter lives. Dolly the sheep is the famous example; she was confirmed a "clone" due to her shorter telomeres, and died many years before a normal sheep was supposed to.

Could this be used as a clock? Conceptually, like (as you said) from a science-fiction standpoint, it's possible. A mad scientist could organize some sort of non-mutating chromosome that has super-duper long telomeres that also can't be affected by telomerase. Then, counting the telomeres could determine the number of divisions, and the relative age of the cell. Emphasis on the "fiction" side of all of that.

On the flip side, though, there are presently several mad scientists attempting to use telomerase to extend the lives of cells; some believe it to be the closest thing to a "fountain of youth," and it seems to work well with human skin cells. No fiction with that one.
posted by BenzeneChile at 12:33 PM on December 12, 2007

I found a few by googling for "estimating age" "blood sample". Here's one such study: "Estimating age of humans based on telomere shortening." Forensic Science International, Volume 126, Issue 3, Page 197, A. Tsuji

Here's a PPT on the subject that summarizes and refers to various articles.

This refers to a study on the subject in Mississippi crocodiles.
posted by aeschenkarnos at 2:40 PM on December 12, 2007

More broadly speaking, this kind of thing is of interest to geneticists generally but of compelling interest to forensic scientists in law enforcement. I'd ask them, here's a forensic science discussion forum. Try not to sound quite so much like an astrologer. :D
posted by aeschenkarnos at 2:43 PM on December 12, 2007

Try not to sound quite so much like an astrologer.

Sssshhh! I was actually trying to sound like a writer!
posted by christopherious at 3:25 PM on December 12, 2007

langeNU, the problem with your scenario is that DNA replication is semiconservative, so the radioactivity won't be propagated throughout the body.

1) radioactive P introduced

2) DNA replication and division.

3) Now you have two cells, each with one hot strand and one normal

4) P32 removed

5) DNA replication and division.

6) Now you have four cells, two of which have a single hot strand, and a normal
strand, and two of which have all normal strands.

7) After N divisions, you'll have 2^N-1 cells, and only two of them will have any radioactivity. How on earth would you find that original cell and accurately measure the levels of P32? It becomes a needle in a haystack problem.
posted by chrisamiller at 4:11 PM on December 12, 2007

Worse than that, metabolic processes (i.e. eating food and growing up) will naturally bring in freshly labeled P32.
posted by Blazecock Pileon at 4:18 PM on December 12, 2007

langeNU also misses a few other important points about P33 and P32: Both have very short half-lives (on the order of weeks), further exacerbating the needle-in-the-haystack problem of actually finding labeled strands after any reasonable amount of time. Beyond that, both are beta-emitters. Cancer-tastic, for mother and child!

Studies of mutation rates may also be problematic: any mutation that affects all your DNA had to have happened when you were a fertilized egg. Subsequent mutations are only going to affect daughter cells that get the chromosome with the mutation on it - meaning again that when hunting for mutations, you're again looking for a needle in a haystack, since you need to find the mutation hidden somewhere in the chromosomes of the cell AND you need to luck out and find mutant cells. You might be able to do a large sample of cells across the body and compare the genome of each cell against the original genome, but that's a fairly extraordinary amount of work by today's standards, and might or might not yield useful results. That's doubly the case when you take into account the fact that there's still a lot we don't know about DNA repair systems. (n.b. - I mostly study stuff on a smaller biochemical level (DNA damage and repair) so I'm not necessarily up to date on larger and more general studies of somatic mutation rates and whatnot.)

Telomeres are the most probable - but again, there's still a great deal that we don't know.

(Directly addressing the possibilities in your question: exposure to the sun as a way to determine birth date would run into the mutation problems described above, in that you might get a few mutations, but they'd be in a handful of cells, and depending on the baby's DNA repair system might be eradicated pretty quickly anyway. Radioactive labels come closest on the "beats of a clock" front, but run into the problems described above - most labels would be carcinogenic, too long-lived like C14, or too short-lived.)
posted by ubersturm at 4:35 PM on December 12, 2007

Given that a DNA sample normally contains stuff other than plain ol' DNA (ie, is blood, sweat, urine, etc) it may well be possible to use those clues to help determine the approximate age of the person involved. However there is unlikely to be any process in reality that will give an answer without being harder than just plain identifying the individual concerned.

In terms of approximate age, you might be able to get results along the lines of infant, prepubescent, pubescent, young adult, adult, menopausal woman, elderly. You could also discover other identifying factors if they are particularly unusual (gross obesity and Down's syndrome spring to mind).
posted by aeschenkarnos at 5:12 PM on December 12, 2007

Don't forget to take into account that freckly and mole-y adults may have naturally longer-than-usual telomeres, which would make them seem genetically younger than their chronological age would otherwise suggest.
posted by Asparagirl at 7:09 PM on December 12, 2007