Researching a novel - anyone know anything about DNA

[medusa]

If a precise half-match (7 of 14 loci tested, for example) was found then that would show up as a familial match, but it couldn't be said to be 'we think it's your father'. Quite apart from the fact that there's a confidentiality issue there, the same 7 loci would be likely to be shared with a child (so the person's child as well as their father- think illegitimate children) and could also completely by chance be shared with siblings, including half-siblings.

 

What you have to remember is that there's statistically a much higher chance of a random match occurring with only 7 loci being tested than with a higher number, so the odds of a random match at 7 loci to someone totally unrelated would have to be ruled out with further family tests.

 

If the original test was at all degraded (such as evidence from some time ago) then it's entirely possible that the number of loci which can be tested and typed is much lower than normal and this would generate only a partial profile. Every element which brings down the number of loci which could be tested for comparison brings down the statistical likelihood of getting a single unique match of any description. Sometimes the viable DNA for testing is so degraded that there's no chance of getting odds high enough to make a proper match using standard DNA testing.

 

In these cases tests such as mitochondrial DNA are often used, but that wouldn't help at all with catching a man through his son, since mitochondrial DNA is passed down through the maternal line, so the son's DNA would be totally different from the father's.

 

If you were to try to use mitochondrial DNA to back up a standard DNA test done when there's degraded DNA for test, it would of course bring up anomalous results and would require further testing.

 

When obtaining a warrant for DNA samples I would imagine that more proof other than a partial DNA match would be required in any case.

Edited May 4, 2009 by medusa
typo

[19SazzaLou91]

I wonder if anyone can help me with some research for my novel – there’s so much expertise on the Forum that I thought I’d start here! (I’ve already done quite a bit of research, but every re-draft raises more questions!)

 

In my story, a man’s DNA is fed into the national database after he’s arrested for a minor offence. The database throws up a ‘familial link’ to the profile of the unknown perpetrator of crimes dating back to the 70s. The police think the offender must be my character’s father. I know that my character’s DNA would be similar enough to that of his father to throw up a partial match. But how similar would it be to the profiles of, for example:

 

His father’s brother

His father’s son by another woman (not our man’s mum)

His mother’s brother

 

Basically, if the police came saying ‘your DNA profile is very similar to one we have on file, we think it’s your father’. Could my man reasonably argue that it’s as likely to be an uncle or half-brother?

 

Would be very grateful for any help!

 

 

Your characters DNA (after having done a DNA test) will have similar patterns to your characters father. Ie: they may have bold lines in the same place, or patterns of similar traces.

However, the similarity to his step-brother would be different. The patterns may be similar, but enough to know that he is not closely related to the character. (They may share the patterns from the fathers DNA). However, to get a full match, you would need your characters mothers DNA in order to argue that.

For an uncle, similarities would be fairly uncommon. You would be able to tell it is not the same DNA line. (if you know what I mean).

 

If you ignore the flashing thing (it was the best thing I could find to explain)

:http://www.odec.ca/projects/2004/mcgo4s0/public_html/t5/four_human_DNA_patterns.gif

You have four different DNA test reults. If you imagine the one with the flashing cross is your character, and the other three were the three you mentioned, you can tell which are closely related, and which aren't. (They go in columns).

[selliot]

Thank you so much everyone for these fantastically helpful replies - you are a clever and knowledgable lot!! It's such a complicated area, and I find I just think I've got it right when something else comes up to throw it all in the air again. Anyway, you've clarified some things for me, and I now have several more avenues to explore. Thank you all you so much! When (if!) the book is eventually published, you helpful forummers will definitely get a mention in the acknowledgements. Watch this space!

[Crayfish]

I wonder if anyone can help me with some research for my novel – there’s so much expertise on the Forum that I thought I’d start here! (I’ve already done quite a bit of research, but every re-draft raises more questions!)

 

In my story, a man’s DNA is fed into the national database after he’s arrested for a minor offence. The database throws up a ‘familial link’ to the profile of the unknown perpetrator of crimes dating back to the 70s. The police think the offender must be my character’s father. I know that my character’s DNA would be similar enough to that of his father to throw up a partial match. But how similar would it be to the profiles of, for example:

 

His father’s brother

His father’s son by another woman (not our man’s mum)

His mother’s brother

 

Basically, if the police came saying ‘your DNA profile is very similar to one we have on file, we think it’s your father’. Could my man reasonably argue that it’s as likely to be an uncle or half-brother?

 

Would be very grateful for any help!

 

As I understand it:

 

It should be 50% to the father and about 25% to each of the others... approximately, in theory, over the whole genome. The loci tested for forensic purposes (short tandem repeats) may be slightly more variable.

 

The DNA profile tests carried out for database / 'fingerprinting' purposes are not the same ones that would be carried out for genealogical purposes or paternity tests. Paternity tests generally use a set of Y-chromosomal loci.

 

Alone, the fingerprinting data could suggest relatedness but would not be sufficient evidence without further tests, and would be unlikely to be flagged as similar on the database without prior suspicion, as with this battery of tests, a few people on the database may well be 50% similar without being closely related.

 

If they already suspected his father and did notice this relatedness for some reason, more detailed tests would be required to confirm it (assuming that reasonably intact biological material from the father is still available).

 

However, fiction often takes small scientific liberties and not many people would know the details of forensic DNA testing.

 

Here's wikipedia's take on the matter, which includes links to various sources http://en.wikipedia.org/wiki/Genetic_fingerprinting#cite_note-7

 

Additionally, I don't know the setting of your novel, but sequencing methods improve quickly and it will probably be cost effective to gather much more detailed data in 5-20 years time (I guess this wouldn't affect the original tests on his father, though).

Edited May 5, 2009 by Crayfish

[selliot]

As I understand it:

 

It should be 50% to the father and about 25% to each of the others... approximately, in theory, over the whole genome. The loci tested for forensic purposes (short tandem repeats) may be slightly more variable.

 

The DNA profile tests carried out for database / 'fingerprinting' purposes are not the same ones that would be carried out for genealogical purposes or paternity tests. Paternity tests generally use a set of Y-chromosomal loci.

 

Alone, the fingerprinting data could suggest relatedness but would not be sufficient evidence without further tests, and would be unlikely to be flagged as similar on the database without prior suspicion, as with this battery of tests, a few people on the database may well be 50% similar without being closely related.

 

If they already suspected his father and did notice this relatedness for some reason, more detailed tests would be required to confirm it (assuming that reasonably intact biological material from the father is still available).

 

However, fiction often takes small scientific liberties and not many people would know the details of forensic DNA testing.

 

Here's wikipedia's take on the matter, which includes links to various sources http://en.wikipedia.org/wiki/Genetic_fingerprinting#cite_note-7

 

Additionally, I don't know the setting of your novel, but sequencing methods improve quickly and it will probably be cost effective to gather much more detailed data in 5-20 years time (I guess this wouldn't affect the original tests on his father, though).

 

Thanks Crayfish, again, very useful! You're right that not many people would know the details - it's a question of trying to get enough details right to make the 'scientific liberties' plausible. Thanks again.

[esme]

remember there is a huge difference between the dna profile used in forensic testing and the dna pattern as used in the human genome project

 

a dna profile gets produced by cutting up a sample of DNA with enzymes into different sized chunks and this is passed through an electrophoresis gel with bigger bits not travelling as far as smaller bits in the same time, the pattern you get is used as a profile, it's similar to chromatography which you may have done in chemistry at school with an ink blot on filter paper to separate out the component dyes

 

the dna pattern is a list of all the base pairs in sequence and as such is a completely accurate representation of a persons dna

 

you also need to bear in mind that once in the environment DNA degrades over time, light, chemicals and even bacteria will all tend to compromise a sample over time until eventually it is useless

 

and don't forget the confidence limits on the results, no forensic expert will say "this dna profile matches that person exactly" instead they will give a probability of a match of say 1 in 10,000 or 1 in 50,000 and this needs interpreting correctly, most people hear "1 in 50,000" and think there's no chance of a match with anyone else, what this really means is that for every 50,000 people they would expect 1 match and in a country of 65 million that means there should be around 1300 matches and familial groupings will tend to cluster these in specific areas so it's not 1300 over the entire country it's more like 1300 over the entire country with around 1000 within a 10 to 15 mile radius, though I'm guessing on the radius figures as I've seen no research for it

Edited May 5, 2009 by esme

[selliot]

remember there is a huge difference between the dna profile used in forensic testing and the dna pattern as used in the human genome project

 

a dna profile gets produced by cutting up a sample of DNA with enzymes into different sized chunks and this is passed through an electrophoresis gel with bigger bits not travelling as far as smaller bits in the same time, the pattern you get is used as a profile, it's similar to chromatography which you may have done in chemistry at school with an ink blot on filter paper to separate out the component dyes

 

the dna pattern is a list of all the base pairs in sequence and as such is a completely accurate representation of a persons dna

 

you also need to bear in mind that once in the environment DNA degrades over time, light, chemicals and even bacteria will all tend to compromise a sample over time until eventually it is useless

 

and don't forget the confidence limits on the results, no forensic expert will say "this dna profile matches that person exactly" instead they will give a probability of a match of say 1 in 10,000 or 1 in 50,000 and this needs interpreting correctly, most people hear "1 in 50,000" and think there's no chance of a match with anyone else, what this really means is that for every 50,000 people they would expect 1 match and in a country of 65 million that means there should be around 1300 matches and familial groupings will tend to cluster these in specific areas so it's not 1300 over the entire country it's more like 1300 over the entire country with around 1000 within a 10 to 15 mile radius, though I'm guessing on the radius figures as I've seen no research for it

 

Thanks for this, esme. I take your point! If I understand it correctly, the dna similarity would only form one part of the evidence in a case like this? I've spoken to a few police officers about the storyline, and they seem to think it was plausible, but that was before I was being asked for more specific details! Most of the idea for this particular storyline in my novel came from the case of the dearne valley shoe rapist, who was caught after his sister committed a minor offence and her dna showed a partial match. Would it be more believable if the police were interviewing my character as one of, say, 30-40 people with a similar dna profile to the one created from the original crime scene?

[Dave59]

Hi Selliott

 

Regarding your question.

 

If the DNA test showed up a familial link, then it would be just that - familial. The police would be able to say that this sample links to a general family group, as in the case of the Dearne Valley Rapist. In some recorded cases however, the police have generated a group of up to 100 persons having a similar profile, this profile was then used not as evidence in a trial, but as a basis for further investigation leading to a conviction.

 

Medusa raises an interesting point regarding the loci. In the early days of DNA testing, when the forensics only tested a low number of loci, there were a number of false positives. Perhaps most interesting was the case of a disabled, man with extremely limited mobility who was suspected of being involved in a burglary that included climbing through a window. Perhapsd this might be an interesting addition to your story?

 

I have included a link with relevant research to the stories I mentioned.

 

With my regards

 

DD

 

http://www.genewatch.org/uploads/f03c6d66a9b354535738483c1c3d49e4/NationalDNADatabase.pdf

[Hopman]

If you have a library card you can access information via the Council/library website 24 hour service.

There will be sources linked to the Credo Reference service.

[selliot]

Hi Selliott

 

Regarding your question.

 

If the DNA test showed up a familial link, then it would be just that - familial. The police would be able to say that this sample links to a general family group, as in the case of the Dearne Valley Rapist. In some recorded cases however, the police have generated a group of up to 100 persons having a similar profile, this profile was then used not as evidence in a trial, but as a basis for further investigation leading to a conviction.

 

Medusa raises an interesting point regarding the loci. In the early days of DNA testing, when the forensics only tested a low number of loci, there were a number of false positives. Perhaps most interesting was the case of a disabled, man with extremely limited mobility who was suspected of being involved in a burglary that included climbing through a window. Perhapsd this might be an interesting addition to your story?

 

I have included a link with relevant research to the stories I mentioned.

 

With my regards

 

DD

 

http://www.genewatch.org/uploads/f03c6d66a9b354535738483c1c3d49e4/NationalDNADatabase.pdf

 

Thanks Dave59 - I've pm'd you