Hi Andrew Richards

I'll answer Question 3 first, as it will give some background for the other two.

When you are analysing a sample using droplet digital PCR, the raw data output is number of droplets that were positive (i.e. contained one or more target molecule) and the total number of droplets (positive and negative) that were in the expected size range. Both of these numbers are integers because they are a count of number of droplets.

The copy number that is calculated by QuantaSoft is calculated like this:

1) The average copies of target per droplet (we call this value Lambda) is calculated by applying a Poisson distribution to the number of positive droplets divided by total accepted droplets. This is done to account for the presence of more than one target molecule per positive droplet (which isn't really very likely at the concentration range you are working in, but once you get to higher concentrations is is more likely than not).

This number can be a non-integer because it is an estimate based on the Poisson distribution.

2) The estimated number of target copies per µL is calculated by multiplying the average number of copies per droplet by the droplet volume and converting to µL. Again, since Lambda and the droplet volume are both non-integers, the concentration value is also not an integer.

For Question 2, you could round the numbers if you like bearing in mind that you will be rounding to only one significant digit for some of your results, however at very low copy numbers like this it probably doesn't matter. Please be careful about drawing strong conclusions from very low copy numbers, as the uncertainty from the Poisson distribution will be large compared to your signal.

For Question 1, copy number data is on a linear scale, unlike real-time qPCR data which is on a log scale, so it doesn't make sense to transform it.

The copy number concentration for repeated analysis of the same sample will conform to a normal distribution, however you might need to run four or more replicate ddPCR reactions per sample to be able to see it at very low copy numbers.

From my experience, the variability of sub-sampling the soil and extracting your target pathogen DNA from the soil will be very large compared to the between well variability of the digital PCR so this might be contributing to the difficulty in analysing your data so make sure you have enough replication at both steps to capture that variability and allow you to see if there is a true difference between your treatment groups.

Feel free to comment reply or PM me if you have any other questions.

Regards,

Jacob.

Jacob L. H. Mclaughlin Thank you for the thorough explanation. I did some further reading and have a solution!

How we can convert this copies number into a log value ?