PDA

View Full Version : A GPS in your DNA



GTC
08-19-2012, 07:30 AM
From Science Daily ...

"Prof. Eran Halperin of TAU's Blavatnik School of Computer Science and Department of Molecular Microbiology and Biotechnology, along with a group of researchers from University of California, Los Angeles, are giving new meaning to the term "genetic mapping." Using a probabilistic model of genetic traits for every coordinate on the globe, the researchers have developed a method for determining more precisely the geographical location of a person's ancestral origins."

Full article: http://www.sciencedaily.com/releases/2012/08/120816101039.htm

ilmari
08-19-2012, 07:51 AM
Hmmm,

The link is not working.

Could you tell us the source and perhaps we can find it on our own?

GTC
08-19-2012, 09:25 AM
Not sure what the problem was, but I've re-posted the link now and it works for me.

zaender
08-19-2012, 09:36 AM
I could see it and it is similar to what we did with Polako - combining BGA plots with GPS data. If you are not too admixed, there was a good chance to end up very near! All my wife's four grandparents could be located on a W/E axis at 20 km.

Spark
08-19-2012, 09:53 AM
I could see it and it is similar to what we did with Polako - combining BGA plots with GPS data. If you are not too admixed, there was a good chance to end up very near! All my wife's four grandparents could be located on a N/E axis at 20 km.

Just when I got my hopes up. There goes that. LOL.

scottraveler
08-19-2012, 02:02 PM
From Science Daily ...

new meaning to the term "genetic mapping." Using a probabilistic model of genetic traits for every coordinate on the globe,

Full article: http://www.sciencedaily.com/releases/2012/08/120816101039.htm

I love science reporters: "every coordinate on the globe"? There are only two --- latitude and longitude. Perhaps they meant "every location on the globe"?

AJL
08-19-2012, 04:41 PM
Dr. Doug McDonald is really the pioneer in this but now anyone with a computer can do this to some extent. It also "works" for admixed people with the proviso that you treat it as geographical weighted averaging of your ancestries and not a single literal spot.

For example, with different studies and methods my average is usually given as somewhere between Moldova and southeastern France, always around the 47th parallel, but there are multiple ways to arrive at this average, and the east-west confidence is much weaker than the north-south. On DNA-Forums, I posted about how to interpret Population Finder (which McDonald designed and Family Tree DNA later tweaked themselves) non-literally for admixed people. Unfortunately I don't seem to have the image any more, but the basic principle is that you can use GPS coordinates as numbers, average numbers, and arrive at a spot on the map. You can also do vice versa if you know one parent's location but not the other, with the proviso that if they are admixed themselves, the location coulod be very rough since you may not inherit their ancestral popualtions evenly.

Seabass
08-20-2012, 01:09 AM
Something I would have to see to believe. I would not think mutation rates are rapid or significant enough over the space of one ancestral generation and even if they were, how could you determine whether this was spurred on by an environmental factor very unique to a city or town? I highly doubt a professional geneticist as of yet (or maybe ever) could see say a nomadic individuals genealogy written in their DNA. If its hard enough discovering regional genetic differences amongst inhabitants of Britain and Iberia, despite hundreds having now been sampled, I highly doubt one can pin-point variation at an intra-county or district level.

ilmari
08-20-2012, 09:15 AM
My data was pin pointed spot on.

MJost
08-20-2012, 01:33 PM
Dr. Doug McDonald is really the pioneer in this but now anyone with a computer can do this to some extent. It also "works" for admixed people with the proviso that you treat it as geographical weighted averaging of your ancestries and not a single literal spot.

For example, with different studies and methods my average is usually given as somewhere between Moldova and southeastern France, always around the 47th parallel, but there are multiple ways to arrive at this average, and the east-west confidence is much weaker than the north-south. On DNA-Forums, I posted about how to interpret Population Finder (which McDonald designed and Family Tree DNA later tweaked themselves) non-literally for admixed people. Unfortunately I don't seem to have the image any more, but the basic principle is that you can use GPS coordinates as numbers, average numbers, and arrive at a spot on the map. You can also do vice versa if you know one parent's location but not the other, with the proviso that if they are admixed themselves, the location coulod be very rough since you may not inherit their ancestral popualtions evenly.

AJ,

You may have seen this but EuroGenes Project produced a Plot (found here: http://bga101.blogspot.com.au/2012/05/genetic-map-of-west-eurasia-with.html ). Polako blogged that he chose another algorithm, which has just appeared online, called SPatial Ancestry Analysis (SPA).

Here is my split which is has a Northwest and Southeast position either side the British Core showing my Dutch maternal and my unknown Irish,Scottish, Isle of Man and English paternal YDNA based. The end of the Red line identifieds each my dual plots: https://docs.google.com/file/d/0By9Y3jb2fORNNnVqNkJsT0dRTVU/edit

Davidski states that the SPADUAL

"SPA offers a mode in which it assumes samples are of mixed ancestry, and thus can divide their genomes into two halves (represented by two sets of PC coordinates). One way to look at it is that each half represents a parent, but that would only be accurate for people who really are of mixed origin, with their parents coming from two clearly distinct groups in terms of genetic structure."

MJost

AJL
08-24-2012, 11:12 PM
Thanks, looks like interesting software.

MJost
08-25-2012, 03:48 AM
I think so too. In my case, I always plotted Northwest, north and northeastern Euro on different plots depending on who's zombie admixtures are used and what they are looking for deep or recent ancestry. I used the Gedmatch version of MDLP World22 and had interesting results. AJ have you tried it yet?

AJL
08-25-2012, 06:15 AM
Not yet, but perhaps before long.