Digital PCR relies on proportion of positive and negative results in the droplets. What is the guarantee that not all droplets will have atleast one template and thence all droplets will be positive which renders the results invalid. Are there prequired dilution of the sample or some pre-treatment to avoid that or does it happen but rarely or there is guarantee that it would never happen.
Thank you
My lab made a cell dilution worksheet for loading single bacterial cells into alginate droplets. It assumes Poisson distribution. You could probably use it if you know the concentration of template (instead of cells) and the approximate size (radius) of the microdroplet. The worksheet will tell you the proportion of empty vs. occupied microspheres.
Here is the link:
http://sequence.stanford.edu/cgi-bin/Cell_Loading_Worksheet/cell_loading.cgi
My lab made a cell dilution worksheet for loading single bacterial cells into alginate droplets. It assumes Poisson distribution. You could probably use it if you know the concentration of template (instead of cells) and the approximate size (radius) of the microdroplet. The worksheet will tell you the proportion of empty vs. occupied microspheres.
Here is the link:
http://sequence.stanford.edu/cgi-bin/Cell_Loading_Worksheet/cell_loading.cgi
There are some technologies, which use 1-million droplets (Hatch et al., 2011, Lab Chip) up to 10-million droplets (http://raindancetech.com/digital-pcr-tech/). In this case you can be quite sure that there is the expected distribution of single molecules per droplet in your sample without any dilution. However, in case of most system (e.g., QX100/200 by Bio-Rad, OpenArray Platform by Applied Biosystems, Constellation Digital PCR by Formulatrix) one has fewer droplets/partitions (few hundreds up to 50.000). Thus, sample dilution might be a requirement.
Moreover, you have to consider that selected molecules like plasmids of come chained (solved in most cases by enzymatic restriction).
The paper by Kreutz et al., 2011, Anal Chem might give you some further clues.
So I guess the answer is that you to know your samples (concentration) at least a bit.
If you are a user of R you might be interested in our dpcR package (http://cran.r-project.org/web/packages/dpcR/index.html), which offers some function to simulate dPCR experiments.
Just like with qPCR, you should optimize your primers and for each sample type, you should perform a dilution series to be sure you are within the linear range of the instrument. You are correct, if every droplet is positive, then you cannot get a result because the Poisson distribution relies on some droplets being positive and some being negative.
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology
- Molecular Biological Techniques