April 15, 2016 at 03:28AM

Today I Learned: 1) Rain drops are shaped roughly like flattened spheres. Actually, it's a bit more complex, as this video explains between 1:10 and 2:10: http://ift.tt/1qtLRMW. In any case, raindrops are decidedly NOT teardrop-shaped. What things *are* naturally shaped like teardrops, other than Rupert's drops and similarly-viscous things? Thanks to Chigozie Nri for this delightful surprise! 2) ...how to design primers for Gibson assembly! In theory, I already knew how to do this, but knowing intellectually how to design a primer is always different than having done it before. Thanks to Andrey Shur for his putting up with my stubbornness and helping plan out some Gibson cloning! 3) ...the difference between Bio-Rad's digital PCR and RainDance Technology's digital PCR. I think RainDance entered the dPCR market first, and they seem to treat dPCR as a Big Thing, like flow cytometry or mass-spec or confocal microscopy. Their machines are super-accurate, even for digital PCR, which they achieve by making on the order of 10 million individual reaction drops (as opposed to Bio-Rad's ~20,000 drops) and recommending that you load so that only about 1% of those drops get DNA. That way, the number of drops with *more* than 1 piece of DNA is really, really, really small, which gives you a better estimate of the actual concentration of your sample. Unfortunately, making 10 million drops of anything is pretty hard, which makes RainDance machines expensive and slow. Bio-Rad, in contrast, seems to be aiming to make dPCR more of a Routine Thing, more on par with qPCR or plate assaying. Bio-Rad's machine only makes about 20,000 drops per experiment, which can be done with much simpler, cheaper (~1/2 the capital cost) technology than RainDance's. It's also much faster to make and analyze 20,000 drops than to make 10 million -- the full RainDance pipeline takes on the order of 8 hours, while most of the time for a Bio-Rad run is the ~2 hour PCR incubation time. Bio-Rad also has packaged and marketed their dPCR machines and reagents in a more user-friendly, black-box, plug-and-play sort of way, which is both good and bad. On the one hand, everything's ready to go right out of the box, and it's pretty simple to use. On the other hand, RainDance also sells nice kits, and their machines are a bit more flexible in terms of substituting your own PCR chemistry in the droplets. All in all, sounds like Bio-Rad is by far the better system for most applications. Bonus fact: It's "digital PCR" or "dPCR", not "digital qPCR" or "dqPCR". Thanks Erik Jue!