So – I keep getting asked about this. I am looking for how people are doing DNA extractions from 1000s of samples.
Some questions
1. What are the cheapest service providers out there for doing DNA extractions from 1000s of samples?
2. What would be the best way to do this oneself? That is, if you were setting up a new lab or project, how would you set this up to do this cheaply? Robotics? Microfluidics?
3. How are you doing this now?
We use MoBio PowerSoil 96-well extraction kits – not the cheapest, but consistently yield the highest-quality DNA from a wide range of environmental samples. We have also used direct PCR for certain projects (but usually better just to spend the time/$ to do the extractions up front). Direct PCR details here: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0044563
I would be wary of out-sourcing the DNA extractions or trying to do them on the cheap. Samples are usually way too important and it is actually quite easy to contaminate samples or get low-quality DNA. Likewise, we’ve found that DNA extractions with a liquid handling robot can be problematic and don’t actually save much time. Better to spend the time/effort/$ and do it right the first time.
We use mobio powermag as well. Not cheap and I disagree with Noah about the quality and consistency (works decently on fecal, inconsistantly on soil depending on the exact soil, and very poorly on tissue). I have had better results with a variety of soils using MP Bio fastspin kits but have yet to try out their 96 well format kit. It’s even more expensive than mobio.
I think if someone has thousands of the same sample type it’s likely worth it to develop a homebrew kit that is based on chemical extraction like CTAB/SDS that’s tuned for their sample type. The commercial kits weight speed and ease of use over optimal extraction
What are you wishing to extract DNA from and for what purpose? We have a method in my lab for extracting DNA from BACs for Illumina Nextera sequencing that could probably be applied to other applications with small modifications. All in, culturing, extraction, library construction, sequencing <$5 per BAC. Everything done in 384 format.
Jim – Sounds very interesting! How robust is it and do you end up getting much of the bacterial genome when you Nextera sequence as well? (One man’s contaminant is another man’s gold…)
Failure rate <10%, costed for ~15% I think. We treat the host to some DNase; off the top of my head I can't remember what percentage of sequence turns out to be host but it is small enough to easily screen out. We're using this method to crunch hexaploid wheat and other awkward plant genomes (https://pag.confex.com/pag/xxii/webprogram/Paper12096.html) but if bacteria is your thing then I'm pretty sure it could work with a few minor changes. My colleagues have done some exploratory work in that direction already.
We use a modified bead beating protocol with autoclaved small glass beads in a fast prep and then prot k and lysozyme follows by phenol chloroform. You can buy a ingredients in bulk cheaply and it efficiently loses gram positives and eukaryote tissue. I’d guess it’s around $0.5 each but that’s back of envelope
See Storify here:
I am using a method similar to Nick Loman for microbiome samples. I order garnet beads like those used in powersoil and make bead beating tubes. Bead beating is in 1% SDS. This is precipitated with 5M potassium acetate. Then, SPRI beads at a 1:1 or 0.5:1 ratio (homebrew) are used to clean up. Because they are homebrew SPRI the cost for large samples of 500-800ul lysate is well under 50 cents compared to PowerSoil’s $4.50. Non-toxic and robot friendly, works great on most samples ( like leaves ) but some of the inhibitors in soil can still be a problem.
We store samples in Lysing Matrix E tubes. When read for extraxtion, bead beating followed by use of Maxwell 16 using Blood LEV kit. Can process 12 samples at a time (typically we do 9 samples + 3 negative controls). Takes about 45 minutes for 12 samples – with the last 30 minutes being unsupervised (Maxwell16 does its thing), which gives time to prep (I.e. bead beating) more samples for the next run. Costs were not too different from using other kit based methods. We used to do phenol:chloroform based extraction – which was great, but started running into issues of co-concentration of PCR inhibitors.
we use 96-well DNA extraction kit (zymo) and a liquid handler. We do two plates at one time and it takes 6 hours. Not sure whether this is a very cheap way, but it just takes a week for 1000 samples, which works for our purposes.