Polymerase chair reaction results

Faith

I'm doing a problem for uni that involves looking at the results of PCR and analysing them. My question is, how many bands need to be in common before it becomes important. For example, I'm looking at two sets of results, and they have two bands in common, out of 12. Is that just coincidence, or does it imply they're related in some way?

Secondly, part of the question deals with a hypothetical forensic investigation. Skin cells were found that have 5 bands in common with a suspect. Would it still be possible that the skin cells belong to the suspect, or is it more likely belonging to a relation (parent or child) of the suspect?

Finally, why do 6 primer pairs usually generate twelve bands in PCR?

Thanks for any help!

Marksie

Faith wrote:

I'm doing a problem for uni that involves looking at the results of PCR and analysing them. My question is, how many bands need to be in common before it becomes important. For example, I'm looking at two sets of results, and they have two bands in common, out of 12. Is that just coincidence, or does it imply they're related in some way?

Secondly, part of the question deals with a hypothetical forensic investigation. Skin cells were found that have 5 bands in common with a suspect. Would it still be possible that the skin cells belong to the suspect, or is it more likely belonging to a relation (parent or child) of the suspect?

Finally, why do 6 primer pairs usually generate twelve bands in PCR?

Thanks for any help!

OK is this forensic analysis... microsatellites?

Faith

They are involved, yup.

Marksie

Faith wrote:

I'm doing a problem for uni that involves looking at the results of PCR and analysing them. My question is, how many bands need to be in common before it becomes important. For example, I'm looking at two sets of results, and they have two bands in common, out of 12. Is that just coincidence, or does it imply they're related in some way?!

This depend on what you are looking for.
I am guessing parentaage analysis or suspect identification.
using either microsatellites or PCR rflp??
Could you be more specific otherwise?.

Think where the nucleotide repeats are that you are using for parentage or forensic analysis and what you are trying to prove, and more importantly their hypervariability.

I assume its a multiplex as you have two tests each with a maximum of six possible bands...yes?
thats three sets of primers

In forensic analysis you would look at as many primer pairs as possible. How Many would be statistically significant (assuming your sample isn't cross contaminated)... you would probably get away with having one of the twelve not matching. You are looking at a mutation in that event.

2/12 is not a match in forensics
IN parentage analysis you get roughly 50/50 from mother and father, in multilocus minisatellites this uquates to sharing bands with the offspring at about 62.5% ( the avergad bandshare between unrelated is about 25%).
Using single locus minis or microsatellites its not too different
2/12 is not a match in parenatge analysis.

Statistics: you can calculate, from dna databases tha probability of a particular size band in microsatellites occurring by chance. shall we say 0.02? for band one, 0.05 for band 2 and 0.001 for band three.
If you get a match on band one the chances of that are 0.02.
if you get two matching as well its 0.02 X 0.05
if you get all three its 0.02 X 0.05 X 0.001.
Well if you can get all 12 matching, this gives you your large probablilties in forensic cases.

2/12 coincidence .

Faith wrote:

Secondly, part of the question deals with a hypothetical forensic investigation. Skin cells were found that have 5 bands in common with a suspect. Would it still be possible that the skin cells belong to the suspect, or is it more likely belonging to a relation (parent or child) of the suspect?!

See the answer above... do you have to try and guess the parentage type..i.e parent or child?

Faith wrote:

Finally, why do 6 primer pairs usually generate twelve bands in PCR?

Thanks for any help!

What is a microsatellite: a specific hyper variable locus one on each of the chromosome PAIRS.
Thus you will have a maximum of two (heterozygous).

If you only get one you have either a) a mutation which stops pcr or, more likely b) it is the same in both chromosome pairs (homozygous).

these reagion are hypervariable for a reason.

I will let you work around that information now ...
but it gives you teh bare bones

Faith

Nice one, Mark, that's a huge help!

This depend on what you are looking for.
I am guessing parentaage analysis or suspect identification.
using either microsatellites or PCR rflp??
Could you be more specific otherwise?.

Suspect identification, using microsatellites. It's basically just a situation where someone has been murdered, and PCR has been carried out of his DNA, his wifes, hair found at the scene and skin cells found under his fingernails. Also under investigation are two suspects.

I'll have a go now and see how I get on! Cheers!

Jimoslimos

OK, tis been a while since I did any PCR work but I'll give my thoughts.

Faith wrote:

why do 6 primer pairs usually generate twelve bands in PCR

Is there much separation between the bands. What size bp are they? Is one band of much greater intensity on the gel? Primer pairs should be that - a pair (equal size) that run as one band but its not uncommon to get two or sometimes more bands. I'm no expert but I'd imagine it might be possible that one of the pairs could have been cleaved in some way resulting in a band of lower size. Also the primer could have attached itself to part of the genomic DNA - resulting in another band of different size.

Faith wrote:

how many bands need to be in common before it becomes important. For example, I'm looking at two sets of results, and they have two bands in common, out of 12. Is that just coincidence, or does it imply they're related in some way?

Uh I'm unsure on this one. I'd suppose a lot would depend on what region of the genome you are looking at - highly conserved or not.

Faith wrote:

Secondly, part of the question deals with a hypothetical forensic investigation. Skin cells were found that have 5 bands in common with a suspect. Would it still be possible that the skin cells belong to the suspect, or is it more likely belonging to a relation (parent or child) of the suspect?

No I'd imagine that it could still very possibly belong to the suspect, remember that DNA found in a crime scene would be degraded to some extent so 5/12 is still a good match.

Marksie

Jimoslimos wrote:

Is there much separation between the bands. What size bp are they? Is one band of much greater intensity on the gel? Primer pairs should be that - a pair (equal size) that run as one band but its not uncommon to get two or sometimes more bands. I'm no expert but I'd imagine it might be possible that one of the pairs could have been cleaved in some way resulting in a band of lower size. Also the primer could have attached itself to part of the genomic DNA - resulting in another band of different size.

primer pairs are artificially created pieces of DNA roughly 30-40bp pairs in length. Each primer binds to one of the strands in double stranded dna, the area of interest lies in between these and amplification spans the gap in between Vi repeat cycles of DNA denaturation, primer annealing and extension using DNA polymerase.
If either a primer binding site or the area in between is cleaved by degradation it simply does not amplify. Depending upon the degree of degradation you may get a faint amplicon or none at all.
Microsatellites are generally between one and a few hundred base pairs in length and are therefore pretty resistant to degradation on the whole
Mis priming as you describe occurs for several reasons, but any band will occur outside the range expected for that microsattelite locus. however, in forensics analysis these are pretty well defined tests and this should not actually occur.
Further, multiplexed PCR tests usually use fluorescent labels on one or the other primer (hex, FAM, TAMRA, TET) this means that multiple tests can be performed at any given times, the results then being read with a laser and giving peaks of fluorescence. The inclusion of a standard of known size bands within the same well or lane allows the exremely accurate measurement of size and removes any gel deformities which may yield to slightlu different running conditions.

Jimoslimos wrote:

Uh I'm unsure on this one. I'd suppose a lot would depend on what region of the genome you are looking at - highly conserved or not.

No, not in microsattelite analysis. The whole idea behind dna fingerprinting and profileing is that there is massive variability between individuals but uniqueness to AN individual.
A micosattelite is a repeated sequence. They are also known a STRs or simple tandem repeats.
For example, let is consider a tri-nucleotide repeat of compositoin CCC, this may occur 10s or hundreds of time joined together: ccc ccc ccc ccc ccc ccc. Either side is a region where your prmers fit and you amplify in between.
The point is those numbers of ccc are hypervariable and in and given human population database there may be thousands of them giving many diufferent sizes. Which gives an extremely low probability of them matching by chance.
These STR sequences occur in the gaps between coding sequences.. so called junk DNA and are extremely prone to mutation.

Jimoslimos wrote:

No I'd imagine that it could still very possibly belong to the suspect, remember that DNA found in a crime scene would be degraded to some extent so 5/12 is still a good match.

No: Faith you may want to check on partial profiles but if you have a match in 5/12 then where did the other 7 come from? the answer is they are someone elses.
That is the whole basis of you not only proving guilt but also proving innocence .
If you obtained a partial profile then you would only have 5 bands from 5 analyses. BUT i am not sure if this occurrs, but it would be extremely dodgy to cal it a match.

http://en.wikipedia.org/wiki/Microsatellite

Jimoslimos

Hi marksuttonie,

Thanks for correcting my errors. Its been a few years since I did any PCR work and even then it was pretty basic. This has interested me and I have a few questions of my own,

marksuttonie wrote:

No, not in microsattelite analysis. The whole idea behind dna fingerprinting and profileing is that there is massive variability between individuals but uniqueness to AN individual.

Might be completely off-topic but I think I read somewhere that the degree of genetic variation differs between ethnic groups. Not sure if this results in a statistically greater chance of a mismatch

My original point about examining highly conserved regions of the genome stands.......if we're not ruling out the family dog as a suspect:D

marksuttonie wrote:

but if you have a match in 5/12 then where did the other 7 come from? the answer is they are someone elses.

I'm sure I've read that 5/12 or 5/13 is enough to proceed with a conviction. Maybe a statistician could explain the odds. I'll see if I can find the link and reasoning behind

Marksie

Jimoslimos wrote:

Thanks for correcting my errors. Its been a few years since I did any PCR work and even then it was pretty basic. This has interested me and I have a few questions of my own

No problem.. i have about 18 years experience in forensics, paternity testing, plant health and disease, GM analysis and animal genetics. Just to give you some bbackground.

Jimoslimos wrote:

Might be completely off-topic but I think I read somewhere that the degree of genetic variation differs between ethnic groups. Not sure if this results in a statistically greater chance of a mismatch

No its not off topic and you are correct, there were databases created to investigate allele frequency (the bands are termed alleles for forensic purposes). In the very early days using minisatellites, the castro case was one of the first that was challenged... there were several things wrong with it, but i believe..and i am going from memory, you can check it up... that one point was a band that had been prescribed as having a high allele frequency actually was very common between unrelated individuals in the peurto rican ethnic group which castro came from.
http://www.time.com/time/magazine/article/0,9171,957855,00.html?promoid=googlep
What it does is not give you a mismatch but lowers the probabilities of coming from a different unrelated individual.
I will exaggerate here if a particular allele size at a particular locus in a caucasion population occurs at a frequency of 2% then you have a chance of 1: 50 that a random individual will have it. If shall we say you have an ethnic group where it occurs at a frequency of 10% at this allele size, then the chance of any unrelated individual having that particular band at that locus of that size drops to one in ten.
It is NOT a mismatch.
I remember a case in court being thrown out for differing statistical analysis. Two different databases were used, covering the same ethnic population. The first gave a 1: 10 MILLION chance that it was from an unrelated individual. The second gave a 1: 30 MILLION chance. The judge..in a fit of pique, said that the evidence was flawed because that stats were not the same. BUT it was still high probabilities no matter which way you look at it. Now the population data has been pooled for the uk as a whole.
i think certain populations didnt vary from the mean population anyways, some did.

Jimoslimos wrote:

My original point about examining highly conserved regions of the genome stands.......if we're not ruling out the family dog as a suspect:D

Not for parentage or forensics. highly Conserved regions are just that. Now speciation by DNA is another matter... mitochondrial DNA MtDNA uses highly conserved primer sites, but the region in between is variable. This variability changes with time..and is almost entirely due to mutation. THe theory being as mitochondrial are in sperm tails they are lost when the sperm fertilizes th egg.. Nor quite true but it will suffice.
NOW... you can sequence or use restriction enzymes to lok at the exact composition between the highly conserved regions and his can give you the species in question. It can also be used to examine phylogenetic trees.
It has been uses in human genetcis... extraction fromk mummies being one, and determining that the woman claiming to be princess anastasia was another... peter gill did that one .

Jimoslimos wrote:

I'm sure I've read that 5/12 or 5/13 is enough to proceed with a conviction. Maybe a statistician could explain the odds. I'll see if I can find the link and reasoning behind

lets assume the situation is a simple one suspect system.. the 5/13 you are talking about isn't correct. As again you have bands which are present which MUST have come from someone else.. no argument.
What you may be thinking of is if you have 5 perfectly matching bands and no others. The stats are then all about the frequency of alleles and you can multiply those frequencies together..in that instance, 5 could give you a unique profile.

John

Great thread! I miss doing PCRs.

Marksie

John wrote:

Great thread! I miss doing PCRs.

Well if you want real fun try real-time pcr, the reliable way to quantify the amount of amplification product .

Actually i am enjoying it too..its nice to refresh the memory and realise you have still got it.

GuanYin

Noone is using conventionals anymore and most of the better journals don't even accept them anymore.

Faith

psi wrote:

Noone is using conventionals anymore and most of the better journals don't even accept them anymore.

Conventionals?

GuanYin

Conventional PCRs. Basically it's all moved towards Real-time PCR.

Faith

psi wrote:

Conventional PCRs. Basically it's all moved towards Real-time PCR.

Okay. What's the difference between them? I haven't studied PCR in much detail, outside of the aforementioned problem and a keen interest in CSI.

GuanYin

Conventional PCR's are generally non-quantitative or semi-quantitative. They give you an all or nothing answer as the products are usually a saturated amplification.

Real-time is quantitative. It measures actual product after each amplification up until saturation.

Kevster

Isn't it Polymerase Chain Ractions (Not 'Chair')? I'm not being arrogant; I'm merely curious.

Faith

Yes, that was just a typo on my part.

Marksie

psi wrote:

Noone is using conventionals anymore and most of the better journals don't even accept them anymore.

errm do you have a refrence for that?.

You don't need to quantify in forensics or parentage analysis or even for testing for genetic diseases.

In medicine, RT PCR is incredibly useful (viral load in HIV), but conventional PCR has many applications

We are accredited for screeing by convetional pcr of animal feeds for GM, then we identify and finally quantify by RT-PCR.

The eu official methods for plant health and disease involve conventional pcr but not RT-PCR.

r3nu4l

Also, conventional PCR is used all day every day in cloning experiments. I should know, I've done thousands of 'em

Journals have no problem accepting papers where regular PCR was performed, why perform real-time PCR when you don't need to.

What some journals won't accept is someone trying to quantify their product DNA through use of quantitative DNA ladders and normalising software. They prefer more rigorous methods.

If the quantification method used is not central to the rest of the experiment/result then conventional PCR and measuring product off a ladder/UV spectrometer is just fine imo.

Jimoslimos

marksuttonie wrote:

In medicine, RT PCR is incredibly useful (viral load in HIV), but conventional PCR has many applications

We are accredited for screeing by convetional pcr of animal feeds for GM, then we identify and finally quantify by RT-PCR.

The eu official methods for plant health and disease involve conventional pcr but not RT-PCR.

I was about to question why RT-PCR was not conventional PCR and then realised,

RT = Real time
and
RT = reverse transcriptase
Maybe a new acronym is needed

r3nu4l

I've seen people use 'RTase-PCR' in the past but that was where they were using Real-time and reverse transcriptase PCR in the same study.

GuanYin

r3nu4l wrote:

Also, conventional PCR is used all day every day in cloning experiments. I should know, I've done thousands of 'em

Journals have no problem accepting papers where regular PCR was performed, why perform real-time PCR when you don't need to.

What some journals won't accept is someone trying to quantify their product DNA through use of quantitative DNA ladders and normalising software. They prefer more rigorous methods.

If the quantification method used is not central to the rest of the experiment/result then conventional PCR and measuring product off a ladder/UV spectrometer is just fine imo.

Apologies, I was unclear in my intital post. I meant no labs doing top class research are using conventional PCRs as evidence anymore. Good journals (PNAS, nature, science, JCB etc) have stopped accepting them as sole evidence of transcriptional changes. If you can show me a paper from a top ranking journal in the last year the relies soley on conventional PCR, I'll withdraw my comment.

r3nu4l

psi wrote:

Apologies, I was unclear in my intital post. I meant no labs doing top class research are using conventional PCRs as evidence anymore. Good journals (PNAS, nature, science, JCB etc) have stopped accepting them as sole evidence of transcriptional changes. If you can show me a paper from a top ranking journal in the last year the relies soley on conventional PCR, I'll withdraw my comment.

Sorry psi, I think I'm confused here, I'm not trying to pick holes. You've qualified your original post but I never said anything about PCR being used to show evidence of 'transcriptional change'.

Surely PCR can't be used to show transcriptional change anyway since PCR is at the gene level and transcriptional change is at the protein level? Wouldn't transcriptional change be shown by post-source decay MALDI-TOF Mass spectormertry analysis or N-terminal protein sequencing? Even sequencing a PCR product can only 'predict' trascriptional changes but not 'show' them. I don't think journals have been accepting those predictions for quite a while, they would demnd protein sequencing as well.

Just to clarify my earlier point, PNAS, Nature and Science are all accepting papers that use PCR as a general technique, see the halotype analysis method on the last page of this paper published in January 2007.

GuanYin

r3nu4l wrote:

Sorry psi, I think I'm confused here, I'm not trying to pick holes. You've qualified your original post but I never said anything about PCR being used to show evidence of 'transcriptional change'.

Surely PCR can't be used to show transcriptional change anyway since PCR is at the gene level and transcriptional change is at the protein level? Wouldn't transcriptional change be shown by post-source decay MALDI-TOF Mass spectormertry analysis or N-terminal protein sequencing? Even sequencing a PCR product can only 'predict' trascriptional changes but not 'show' them. I don't think journals have been accepting those predictions for quite a while, they would demnd protein sequencing as well.

You're right of course. I actually didn't quite read the thread as well as I should. I just read the opening post there.

Just to clarify my earlier point, PNAS, Nature and Science are all accepting papers that use PCR as a general technique, see the halotype analysis method on the last page of this paper published in January 2007.

Oh come on - could you go more obscure

r3nu4l

psi wrote:

Oh come on - could you go more obscure

Ah no, I take your point, conventional PCR as a method is simply not sexy enough or rigorous enough to qualify as a main focus for a paper in any top-tier journals. It is accepted as part of much larger experiments but the results aren't something you would focus on as evidence without going a step or two further to back up any major claims.

The technique is well established and is a 'bread-and-butter' lab technique these days prety much like a DNA electrophoresis. A stepping stone to bigger things

GuanYin

r3nu4l wrote:

Ah no, I take your point, conventional PCR as a method is simply not sexy enough or rigorous enough to qualify as a main focus for a paper in any top-tier journals. It is accepted as part of much larger experiments but the results aren't something you would focus on as evidence without going a step or two further to back up any major claims.

You bastard!! You knew what I meant and argued anyway!!!!

Oh for the days when I was like you, before apathy kicked in!

r3nu4l

psi wrote:

You bastard!! You knew what I meant and argued anyway!!!!

^*screams and points* "personal abuse, personal abuse!"

Nah, I wasn't really arguing with you at all, it's just that as we all know there are a lot of people on this forum who may not be qualified to any significant level who might get the impression that conventional PCR is next to useless when it does have a valid purpose. I was basically just highlighting that for those people rather than arguing, I know exactly what you mean, that's all
oh, yes, it was obscure too

I know what you mean, I went through several bouts of apathy during my PhD and then again during a post-doc contract

GuanYin

r3nu4l wrote:

I know what you mean, I went through several bouts of apathy during my PhD and then again during a post-doc contract

Well I meant with boards, but all I'll say is wait til you become a PI and go into administrative hell

r3nu4l

psi wrote:

Well I meant with boards, but all I'll say is wait til you become a PI and go into administrative hell

Ah! Well I now work outside the academic life and for a MedComms Agency. I've worked as a Project Manager and Editorial Manager so administration is now a huge part of my job.