Hi, I have absolutely no experience in soil analysis, but considering you have no information about any species present in the soil, and therefore no reference DNA to go (because you don't know if your treatment will increase or decrease number of bacteria, going with ribosomal RNA won't tell if the difference in final amount of these RNA is due to actual extraction efficiency difference or bacteria death) I would go with a fixed amount of DNA you add to each soil sample before extraction.

You have to make sure that the sequence of this DNA cannot in any way be present in the soil sample (for example designing an oligo sequence from the DNA sequence specific of a eucaryote species absolutely not present in your country).

Add say 10.000 copies of the oligo sequence to each soil sample, then do the extraction. After that do a quick PCR with samples and a standard curve of the oligo sequence (10 - 100 - 1000 - 10.000 - 100.000 copies) and you will have your DNA extraction efficiency.

It don't know if something easier might exist, but considering the material you are using (soil), I cannot think you can use an endogenous (= always present in the soil sample in the same quantity) DNA marker to check your extraction efficiency.

Thanks for rephrasing!

If you are mainly interested in the effect of the soil treatment to the microbial composition (in other words, if you expect the treatment to have an effect on the abundances of particular microbes and their genes), then I would probably not worry to much about that. I would extract all samples using the exact same method and measure the amounts of DNA in each of them, maybe adjust for equimolar amounts, and then continue for the shotgun analysis.

In addition, you could do a qPCR on the 16S rRNA gene, to see what the total amount of ribosomal RNA copies in per mg soil. Note that this is not the same as the total amount of bacteria in your sample, because some bacteria have more 16S copies than others, but it is a pretty good measure. There are several good PCRs published, but our group has described a good quantitative (Taqman) broad-range 16S rDNA PCR in e.g.:

http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0003056

So make sure you weigh the soil sample before extracting, because if you want to relate the found number of 16S copies to mg soil.

The efficiency of DNA extraction is hard to calculate. This might be an important thing to worry about, if you suspect that the treatment of the soil will actually hinder the DNA extraction or that components in one of the soils (but not in another) might contain inhibitors that might be co-extracted and hinder downstream reactions. If that is the case, then, you could do what Alexis suggested: Like Alexis mentioned, you could "spike" your soil with known DNA amounts (human betaglobin ?) and see how many of those copies will be found after extraction.

But if you are treating the soils with something that might not have a big effect on DNA extraction, I would probably not worry about this too much. Most DNA extraction kits, eg. the MoBio Powersoil kit or the QIAamp kits, can handle inhibitors, and they are specifically designed to remove most of them.

I hope this is a bit more helpful - there are tons of publications on microbial soil composition, and I have not done any soil extractions myself, so hopefully other persons can chip in.

Just to complement.

One nice thing about DNA extraction kits based in mechanical lysis, is that they are highly reproducible. The important thing is to treat all samples equally. Applying the same bead-beater strength/time for all samples should give similar DNA yields if the soil is the same.

Excellent answers. Thank you all for your help! I'll work with soil treated with petroleum. Maybe it will complicate the efficiency, but the addition of an amount of DNA is going to help me to evaluate this.

The question itself is intriguing and depends what you mean by 'efficiency'. Is yield or quality the most important thing? Choosing one can mean compromising on the other. Also, you have to consider how representative of the community your extractions actually are, and whether there is a substantial bias against extraction from spore-formers. These can be very difficult to lyse, even with bead-based extractions. Even if using a kit (those mentioned by others above are good), it's worthwhile considering and evaluating additional steps, for your samples. Addition of a known quantity of cells (e.g. sporeformers) to your samples during validation of the extraction protocol (of an organism you will not find in the soil - or one containing a specific marker that you can later screen for in the extract using PCR), will help you gauge how efficient your protocol is.