Jump to content

  • Log in with Facebook Log in with Twitter Log in with Windows Live Log In with Google      Sign In   
  • Create Account

Submit your paper to J Biol Methods today!
Photo
- - - - -

Increasing specificity of allele-specific Tq probes?


  • Please log in to reply
2 replies to this topic

#1 Trof

Trof

    Brain on a stick

  • Global Moderators
  • PipPipPipPipPipPipPipPipPipPip
  • 1,199 posts
109
Excellent

Posted 13 November 2013 - 11:41 AM

HI all,

 

the long story short, I needed a detection assay for a certain mutation, preferably as a commercial kit, and the only cost-worthy for this matter were finally ABI Custom SNP Assays (I know they're for genotyping, not a mutation detection but given the problem set-up, I didn't needed too much sensitivity, either there would be a mutation in majority of cells within sample, or I don't care).

 

Now, if you don't know it designing a custom assay for somethig other than SNP on the ABI pages just suck. No, SUCK. But anyway, I needed an assay to distiguish a one-base deletion.

I also designed other two custom SNP assays this way, for C>T substitutions and they work rather fine. This one does not.

 

ABI SNP assays have two allel-specific Taqman probes with MGB so the point is to get "high VIC-low FAM" for wild-type samples and "probably some VIC - high FAM" on the mutated ones. For two-colored genotyping there is a nice plot, that software then clusters.

 

It should look somehow like this:

11.png
(NTC down in the left corner, mutant standard far right (should be about 100%, but can't be really sure now), wild-type top left)

 

 

It doesn't. It looks like this:

15.png

(they have both high fluorescence for the allele, that should not be there, it even called the "100%" mutant as a heterozygote)

If you draw line there you see both "extremes" are too close. I would have a real difficulty to tell any sample that will fall between 100 % wt and 100% mutant, and if there is one possitive, it surely will.

 

From the look at the amplification plots, it's obvious that in contrast to the first assay, where in the corresponding channels the florescence of the "unmatched" allele reached plateau in much lower fluorescence level (so increasing/decreasing cycles would have no further effect), in this assay they seem to have a similar efficiency to the "matched" allele.
Also, FAM and VIC channels crosstalk, but the graph is designed to take care of that, because it displays fluorescence from both channels in kind of visual ratio, so this cannot be the sole effect of channel crosstalk.

 

It seems to me, that the probes are just not specific enough, and that it's something I can't really affect much. 

 

 

I have definitelly ideas how this could be done really right, and is not, but due to the situation won't be and it's not even worth it really (since it is part of the I will never ever again... project) so I want to get quicky over that or just leave it.

For example, I can't be sure, if my standard is really 100% mutant, it only theoretically should be as it was created by OE-PCR mutagenesis and diluted more than 1000x (no cloning of such useless thing, really..), but of course I'm not sure. But from the graph, it seems in general, that samples have to high fluorescence of the "unmatched" allele, than they should have. (I have no access to patient sample with this mutation to test)
So, since the whole amplicon looks like this:

GGAGAACGGCTGCACATGTTCCGTGTGGGGGGCCTTGTGTTCCACGCCATCGGACAGCTGCT

 

  (yes, this short)
orange - primers, blue - probe(s), red - the C, that get's deleted
I wondered, if the stretch of Gs could do some harm, either to assay and the OE-PCR before, so I'm now running it side by side with 3% and 5% DMSO (still don't have betaine)  to see if it helps anything, but then I have no further ideas. Of course, the main reason could be that the design doesn't work (and they only tested that it somehow runs on wildtype samples) and I can't do anything, but if any of you have some idea how to help it, that's not lenghty I would appreciate it.

Thanks.


Our country has a serious deficiency in lighthouses. I assume the main reason is that we have no sea.

I never trust anything that can't be doubted.

'Normal' is a dryer setting. - Elizabeth Moon


#2 jerryshelly1

jerryshelly1

    Veteran

  • Global Moderators
  • PipPipPipPipPipPipPipPipPipPip
  • 328 posts
29
Excellent

Posted 13 November 2013 - 12:25 PM

I am really confused. You designed this to test whether your deletion was present in a large portion of cells. You are looking at relative fluorescence to see how many cells have the deletion, as compared to a non-deletion containing control? What you see is that your "deletion containing cells" do not have the relative fluorescence that you were hoping for. So, you are looking for an easier method to determine the ratio of cells that contain the deletion?



#3 Trof

Trof

    Brain on a stick

  • Global Moderators
  • PipPipPipPipPipPipPipPipPipPip
  • 1,199 posts
109
Excellent

Posted 13 November 2013 - 01:08 PM

That deletion is relatively common tumor supressor mutation in cancer. I have 70 tumor samples and want to test them for various tumor suppresor mutations, this is one of them. I don't need to quantify them, I just want to know if there is the mutation. Of course due to the nature of the assays (they are designed for SNPs, so they can very specifically distinguish 100% wt, 100% mutated and 50/50, for the same purpose is the analysis module that makes the graph) the sensitivity won't be very high, so I would either found a sample with high percentage of the mutation or not.

 

The reason why this is analysed this way (as all allele specific assays) is there is always some fluorescence in the other channel, due to spectral crosstalk, so instead of any Cts or things like this, this assay is "endpoint" it measures endpoint fluorescence in two different channels and plots them in the graph. Ideally there should be a clear clustering visible in the graph, far apart. My problem was, both "homozygous" (if we were talking genotype) samples are too close together.

I don't want to find other way how to detect mutations there, I need to make this one work better.

 

So far I managed this with DMSO. 

MLL15547-DMSO-optim.png

 If there are no other ideas, I will probably run it with 5% DMSO and only if there are any samples looking a bit like positive, then focus on them more. (so far two other mutations in this gene weren't found in my samples at all, but they clustered pretty nicely, I just have my fingers crossed for the p53, there should be at least some mutated)

 


Our country has a serious deficiency in lighthouses. I assume the main reason is that we have no sea.

I never trust anything that can't be doubted.

'Normal' is a dryer setting. - Elizabeth Moon





Home - About - Terms of Service - Privacy - Contact Us

©1999-2013 Protocol Online, All rights reserved.