DNA TEST . whats Needed ?

jonski

I'm thinking Biology , maybe ......


So , what is needed to do a DNA test on somebody to see if they are related . ie ... I think that girl is my half sister , will a test of her dna and mine show that we are related ?

IrishKnight

Yes, basically a sample from you and a sample from the person in question. To simplify it, if you are half siblings you will share half of your DNA so will have many of the same genetic landmarks. Normally said samples will be from a swab from the inside of the cheek.

EDIT: Yes indeed it would be quarter

darjeeling

IrishKnight wrote: »

Yes, basically a sample from you and a sample from the person in question. To simplify it, if you are half siblings you will share half of your DNA so will have many of the same genetic landmarks. Normally said samples will be from a swab from the inside of the cheek.

Half sibs will share a little over a quarter of their DNA: a quarter through inheritance from the common parent, and a little more as we're all ultimately related somewhere in our history.

If putative half sibs are thought to share the same mother, there's potential for rejecting half sibship if the two individuals don't share the same mitochondrial DNA. Conversely, if the individuals are thought to share a father and both are male, then the discovery of non-sharing of Y chromosomes could be used to reject half sibship.

If Y chromosome / mitochondrial tests don't or can't rule out half sibship, then you'd need to test a decent number of genetic markers and have some good figures on the expected sharing of genetic variation between unrelated individuals in order to distinguish half sibs from random unrelated members of the population. Excluding or confirming parentage is easier, as the child should have inherited one of the parent's two alleles at every single position in the genome.

And one more necessary thing we shouldn't forget is written consent!

IrishKnight

Oh yes of course you are right darjeeling, I thought something was amiss when I was writing that comment...

jonski

Thanks for the replies .

It was just that after a phonecall today to the place that does them here it got a little confusing , the girl on the phone gave the impression that you needed information from the father and that information just isn't available . Without going into too much detail there is no parent left alive to use , only the two girls to see if they are half sisters .

Obviously with full consent .

darjeeling

jonski wrote: »

Thanks for the replies .

It was just that after a phonecall today to the place that does them here it got a little confusing , the girl on the phone gave the impression that you needed information from the father and that information just isn't available . Without going into too much detail there is no parent left alive to use , only the two girls to see if they are half sisters .

Testing for half sibs isn't straightforward, as it's possible that e.g. first cousins could by chance share the same amount of genetic variation as do half sibs. There are commercial companies offering sib tests that aim to reveal whether individuals are full or half sibs, but they come with warnings about results not being conclusive - unlike parentage tests - and point out that test confidence decreases as the relationship tested gets more distant.

The person you phoned was correct to say that having a sample from the common parent would make the situation completely different, as then you could show conclusively that two children were or were not the offspring of the parent, proving that they were or weren't half sibs via that parent. Edit: If you didn't have the common parent's sample but had samples from one or both of the other parents, this would help as you could identify which copy of each marker came from the common parent, giving a more sensitive test.[/Edit]

Having looked at a bit of real genetic data for pairs of (admittedly non-human) half sibs and also less related pairs of individuals, I've seen some less related individuals who do indeed happen to share more variation than some half sibs. In these cases, adding genetic profiles of the parents into the comparison immediately makes it possible to resolve the relationships. This data was generated using a similar approach to that used by 23andMe for humans, relying on lots of genetic markers that are not individually very variable.

As I say, it's possible to rule out the two being half-sisters if you think they had the same mother, and yet you find they have different mitochondrial DNA. Confirming that they are half sisters would be much more difficult, and any commercially available test would carry a fair degree of uncertainty.

jonski

darjeeling wrote: »

As I say, it's possible to rule out the two being half-sisters if you think they had the same mother, and yet you find they have different mitochondrial DNA. Confirming that they are half sisters would be much more difficult, and any commercially available test would carry a fair degree of uncertainty.

Thanks again , that clears up what I was thinking about .