September 28, 2016 at 03:32AM

Today I Learned: 1) ...how high-performance chromatography, like high-performance liquid chromatography (HPLC) and gas chromatography (GC) and other fancy methods like supercritical fluid chromatography (SFC). Here's what I knew: Chromatography is similar, in the most basic principles, to separating components of ink by spotting it on paper. A chromatography setup has a column with some kind of binding material (usually called the solid phase), over which you flow a sample (the liquid or gas phase) containing multiple compounds that you want to separate. Because different components in the liquid phase will bind more or less stongly to the solid phase, they move through the column less or more quickly, respectively. This causes them to come out the other end of the column at different times, so by catching what comes out at different times, you can collect different components. (You can also hook a chromatograph directly to a mass-spec to get really clean characterization of complex mixtures.) Here's what I learned: Chromatography is *not* like an electrophoretic gel, where you just flow a sample through and the more it binds, the longer it takes to come out the other end. Actually, whatever your chemicals-of-interest are are usually calibrated to bind really strongly to your column, at least initially! Your compound of interest, and possibly some other contaminants, will bind to the column, but lots of other stuff will wash through. The way you get the compound(s) of interest off the column is by slowly changing what kind of liquid you wash over the column (the buffer). As the buffer composition changes, it changes how well your compound sticks to the column, and at some critical point, your compound will flow off and you can collect it. For instance, nickel columns are used to collect His-tagged proteins*, because His tags stick to nickel pretty well. If you flow a mix of proteins with some his-tagged proteins (say, a cell lysate containing the tagged protein), then the his-tagged proteins will stick to the nickel column. Once all the excess proteins have washed through, you start to slowly add imidizole to the buffer on the column. His-tags are good at binding to nickel, but imidizole is *really* good at binding to nickel. As the imidizole concentration increases, it eventually starts to kick off the his-tagged proteins, and at some critical point, most of the his-tagged protein goes back into solution and flows off the column to be collected. Thanks to Anders Knight for a solid explanation of chromatography! * a His-tag is a bunch of histidine amino acids tacked onto the end of a protein, usually to make it easier to isolate. 2) The song "Barret's Privateers", which is a truly delightful, if somewhat depressing, classic sea shanty (might I recommend this recording by Stan Rogers? http://ift.tt/2dC1rUE) is actually a modern piece written by the selfsame Stan Rogers in the 70s. I would not have guessed that from listening. 3) Well, Elon Musk just released some new information on how he plans to get people to Mars (and beyond, if travellers are interested). I didn't read too much on the details, but basically it's a giant (*really* giant) two-stage reusable rocket. The rocket first takes a capsule containing crew and cargo to low Earth orbit and then falls back to Earth. It's hitched to a fuel container, launched *back* up into orbit to meet the capsule, and delivers the fuel that will get the capsule to Mars. Then the rocket falls back to Earth a second time to be refitted for another launch. The capsule jets away to Mars, using a solar sail for thrust in addition to the initial fuel consumption (and presumably a decelleration burn at the end? Can any Kerbal players fill me in on this bit?). The reusable rocket makes the trip relatively cheap -- Musk is estimating $500,000 per passenger for the initial flight, with costs potentially falling as low as $100,000 per passenger. The catch is that the initial capital investment required to build the rockets isn't trivial. Musk is estimating $10 billion, and he's not known for making conservative estimates. That's... really, really, *really* outside of SpaceX's price range. It would almost certainly require government money, in a sort of space-race-style project to get people to the moon. In that sense, Musk has just proposed a massive potential public project, rather than a company plan. Now, let's put that in perspective. That $10 billion will be spread out over many years. Musk's most optimistic guess is that we could have men on Mars in 2025. That's the *most optimistic* scenario, but it would spread the investment cost over something like 10 years. Really, though, I would guess that the capital investment will be pretty heavily front-loaded, and I'm also going to guess that it's going to take longer than 10 years, so let's say $2 billion a year for a while. That's about $6 per person per year (there are a lot of humans). How much is that, in terms of other things in our budget? In terms of science, a $2 billion a year is pretty big, but not overwhelmingly so. That's about half the budget of the National Science Foundation, which accounts for a huge fraction of scientific funding in the US. It's only a tenth or so of the budget of the NIH, though, and that's seen as worthwhile, so maybe the NIH could pivot massively and get us into space? But let's be real, science isn't the *first* thing to get cut when creating a new project. That honor belongs to education. The Department of Education received $77.4 billion from the federal government in 2012 (actually up a fair bit since 2010!), so we could get there with a ~3.5% cut to education for a decade, paying no more in taxes. Or, you know, we could cut military spending by a third of a percent.