MADELEINE'S DNA - definitive study.

[Carana]

Professor Dan Kane says it is impossible to see evidence of more than 3 persons in the mix.  To have proof of 5 there would have to be 9 peaks showing at any one locus.  That would be a first!
"Statistical weights of mixed DNA profiles"
https://www.youtube.com/watch?v=daRBTopFA1A&index=5&list=UUjpWy3YdG24iySWTBS4jU2Q

I don't recall him saying that...

[Robittybob1]

I didn't hear him say that...

I might have misheard or misunderstood it.    But they ran some trials and were surprised that the number of alleles was lower than expected.  I'll go through it again and see if I can correct myself.

[Carana]

If at any single locus the results showed 3 alleles then there is a mixed sample of at least 2 people.
If at any single locus the results showed 5 alleles then there is a mixed sample of at least 3 people.

So I'd expect in the results on the Renault to show at least one locus the results showed 5 alleles.   Multiple occurrences of any single loci having 3 alleles does not mean there was more than two people in the mix.

Put it this way I'm still confused by the methods of assessing the results.

I'm thinking about that one.

Yes, 5 alleles at any single locus would show a a mixed sample of at least 3 people.

However, if Lowe had Kate and Gerry's profiles (I'm not certain whether he did or not), depending on how many identical alleles they had in common (e.g. we know that they both had at least TH 9.3 and if that value was one of the ones found in the mix), then a 4th allele would show the presence of at least one other person contributing to the mix.

If, say:

At the TH locus, Kate had a 6, 9.3 and Gerry had 7, 9.3, then an 8 would show a 3rd person (which couldn't belong to Madeleine either).

ETA: If ever Kate and Gerry had a common base pair, e.g. if both had 9.3, 9.3, then it would only take one extra allele value to show the presence of someone else.

[Robittybob1]

Professor Dan Kane says 75%  of the time a 4 person mixtures present themselves as 3 person mixes
https://youtu.be/daRBTopFA1A?list=UUjpWy3YdG24iySWTBS4jU2Q&t=1302



"Statistical weights of mixed DNA profiles"

[Robittybob1]

I'm thinking about that one.

Yes, 5 alleles at any single locus would show a a mixed sample of at least 3 people.

However, if Lowe had Kate and Gerry's profiles (I'm not certain whether he did or not), depending on how many identical alleles they had in common (e.g. we know that they both had at least TH 9.3 and if that value was one of the ones found in the mix), then a 4th allele would show the presence of at least one other person contributing to the mix.

If, say:

At the TH locus, Kate had a 6, 9.3 and Gerry had 7, 9.3, then an 8 would show a 3rd person (which couldn't belong to Madeleine either).

ETA: If ever Kate and Gerry had a common base pair, e.g. if both had 9.3, 9.3, then it would only take one extra allele value to show the presence of someone else.

I think you are correct here but the only thing is if we worked back through Gerry's, Kate's and Madeleine's profiles  we  could work out if we would ever get to 37 different alleles for the three person mixture.  It would be an interesting exercise.

[Carana]

Professor Dan Kane says 75%  of the time a 4 person mixtures present themselves as 3 person mixes
https://youtu.be/daRBTopFA1A?list=UUjpWy3YdG24iySWTBS4jU2Q&t=1302



"Statistical weights of mixed DNA profiles"

Yes, but what you'd said was: "Professor Dan Kane says it is impossible to see evidence of more than 3 persons in the mix.  To have proof of 5 there would have to be 9 peaks showing at any one locus.  That would be a first!"

I'm confused... that doesn't seem to be the same thing.

[Robittybob1]

Yes, but what you'd said was: "Professor Dan Kane says it is impossible to see evidence of more than 3 persons in the mix.  To have proof of 5 there would have to be 9 peaks showing at any one locus.  That would be a first!"

I'm confused... that doesn't seem to be the same thing.

Shall I go back and amend that, I think I misunderstood him on that point.  Sorry.

Do you think a sample would ever have 9 or more alleles at a locus?

[Carana]

I think you are correct here but the only thing is if we worked back through Gerry's, Kate's and Madeleine's profiles  we  could work out if we would ever get to 37 different alleles for the three person mixture.  It would be an interesting exercise.

Yes, it would be interesting, but somewhat an intrusion of privacy. There are 3 other DNA profiles in the files (which is an intrusion into the privacy of at least one named person), but out of interest, I compared them with Madeleine's to see how many alleles these unrelated people had in common with hers. I can't remember offhand now, but I remember being surprised at how many there were.

I also compared random profiles I'd found on the Internet (perhaps 5 or 6), and in one of those I found some guy at the other end of the world who had 9 in common.

[Robittybob1]

Professor Dan Kane says it is impossible to see evidence of more than 3 persons in the mix.  To have proof of 5 there would have to be 9 peaks showing at any one locus.  That would be a first!
"Statistical weights of mixed DNA profiles"
https://www.youtube.com/watch?v=daRBTopFA1A&index=5&list=UUjpWy3YdG24iySWTBS4jU2Q

I was mistaken.

[Carana]

Shall I go back and amend that, I think I misunderstood him on that point.  Sorry.

Do you think a sample would ever have 9 or more alleles at a locus?

I'd have to check, but from (vague) memory, I don't think the range of values at any one locus extends that far.

I think that there could be evidence of 5 contributors in certain cases, though, without needing 9 different allele values at a single locus. E.g., if the profiles of 4 family members were known (and they'd share far more alleles in common than unrelated people and therefore fewer distinct alleles showing) and up pops one or more alleles that are not in any of those profiles.

[Carana]

I was mistaken.

No worries, it happens to all of us. ;)

[Robittybob1]

I'd have to check, but from (vague) memory, I don't think the range of values at any one locus extends that far.

I think that there could be evidence of 5 contributors in certain cases, though, without needing 9 different allele values at a single locus. E.g., if the profiles of 4 family members were known (and they'd share far more alleles in common than unrelated people and therefore fewer distinct alleles showing) and up pops one or more alleles that are not in any of those profiles.

Interesting questions. What is the range of values at any one locus?
4 family members eg mum dad and two kids will have no more alleles than the Mum and Dad on their own. 

Can you have a 3 person mix and only have 37 alleles looking at 10 loci and still have 15 out of 19 alleles matching Madeleine's profile? 
Maybe you could if you mixed 3 closely related members eg Kate, Gerry and Gerry's sister just as an example.  So Madeleine is excluded.  No that wouldn't work either as then there would be 19 out of 19 for Madeleine.  Tricky!
I'd like to look at father, father's sister and a sibling to Madeleine.  Excluding Madeleine.
Or Mother, Mother's mother and a female sibling to Madeleine.  Excluding Madeleine.

[Robittybob1]

Imagine if the DNA result could simply be explained by a three person mixture of the Mother, the Mother's mother and a female sibling to Madeleine.  Excluding Madeleine?  That could be explained by a shopping trip with one of the twins in the Renault.

[Mr Gray]

Imagine if the DNA result could simply be explained by a three person mixture of the Mother, the Mother's mother and a female sibling to Madeleine.  Excluding Madeleine?  That could be explained by a shopping trip with one of the twins in the Renault.

the dna results have been simply explained

[Robittybob1]

the dna results have been simply explained

No they haven't or if you think they have tell us please?

This is the answer from the file: "A complex LCN DNA result which appeared to have originated from at least three people was obtained from cellular material recovered from the luggage compartment section 286C 2007 CRL10 (2) area 2. Within the DNA profile of Madeline McCann there are 20 DNA components represented by 19 peaks on a chart. At one of the areas of DNA we routinely examine Madeleine has inherited the same DNA component from both parents; this appears therefore as 1 peak rather than 2, hence 19 rather than 20. Of these 19 components 15 are present within the result from this item; there are 37 components in total. There are 37 components because there are at least 3 contributors; but there could be up to five contributors. In my opinion therefore this result is too complex for meaningful interpretation/inclusion."

That is not simple.
Who are three people in the mix that could give that result?