Mass defects for natural history

Isotopic mass defects are the topic du jour around here. Those tiny mass differences form the basis of a nice multiplexing scheme for analytical chemists – if you assume that different stable isotopes of the same element share exactly the same chemical reactivity.

That assumption of equal reactivity works well enough for many applications, but there’s an entire field of chemistry dedicated the fact that it isn’t always true: isotopic geochemists live in a world where isotopes react differently. Our knowledge of the ancient environment on Earth — temperature, atmospheric composition, seawater chemistry, and many other climate variables — comes from careful measurements of the average abundance of say carbon-13 vs. carbon-12 in samples of limestone, or from the abundance of deuterium (2H) relative to protium (1H) in carefully extracted samples of fossil biomarkers.

Mass spectrometers for geochemistry have been improving rapidly, just as they have been in proteomics or metabolomics, and now geochemists can measure mass defects too. Their primary question is, what is the internal distribution of heavy isotopes in the molecules of the sample? “I know this methane (CH4) contains X% carbon-13 and Y% deuterium,” says today’s geochemist, “but how much of that deuterium is clumped with the carbon-13 in molecules like 13CDH3, and how much is separate in molecules like 13CH4 and 12CDH3?”.

Late last year, one group reported on their mass spec technique for measuring isotope clumping in methane, and found that clumping in their samples was controlled by thermodynamics. (Yes, thermodynamics has an opinion on how isotopes should clump, just as it does on nearly any question in science.) But mass spec isn’t the only game in town. Another group developed a laser spectroscopy technique to measure isotope clumping in methane, and they found that clumping in their methane samples was kinetically controlled.

The two groups weren’t looking at the same samples, so it could well be that they are both right. But newfound analytical techniques to measure mass defects are going to keep geochemists busy for a long time to come.

Mass defection for massive multiplexing


Twelve molecules with distinct mass defects, as synthesized by Frost et al.

The topic du jour around here is stable isotopes, and in particular the mass defect that stable isotopes exhibit. A quick summary:

  • Isotopes of an element have atoms that differ only in their neutron count.
  • The difference between the mass of an atom and its isotope is an apparent neutron mass for that atom. E.g. m_{^{34}S} - m_{^{32}S} = 1.9958 Da, meaning each “extra” neutron in 34S has a mass of 1.9958/2 = 0.9979 Da.
  • The mass defect is the difference between the apparent neutron mass of various isotopes or elements. In the last post we saw the “extra” neutron in 13C was 1.0034 Da, meaning the carbon-vs-sulfur mass defect is 1.0034 – 0.9979 = 0.0055 Da or 5.5 millidaltons.

The natural follow-up question is “OK, who cares?” Well, modern mass spectrometers can resolve these very tiny differences. And that ability can be put to some very interesting ends.

Continue reading

Chemistry: Not quite so elementary


Harold Urey, father of isotopic chemistry. Image from Wikipedia.

Chemistry, like many of the sciences, is laden with a horribly confusing vocabulary that persists for merely historical reasons. For example, atoms are not indivisible even though atom comes from an ancient Greek word atomos meaning “indivisible”. “Free” energy is not free, at least not as most speakers of modern English would understand the word.

And so it is with the chemical “elements”: they aren’t elementary. Take carbon for example. This supposedly elementary substance can be chemically separated into (at least) two types, each with a different reactivity. I’m talking, of course, about isotopes. The heavier type of carbon is carbon-13 or 13C for short, and is made of atoms that have six protons and seven neutrons in its nucleus. The lighter, more prevalent type is carbon-12 (12C), with the same number of protons, but only six neutrons instead of seven. Admittedly, the differences in chemical reactivity of 13C and 12C is very slight — entirely negligible for most practical purposes. So in a sense, calling “carbon” an “element” is only a slightly misleading, but there are certainly times when the difference can be very important.

Don’t just take my word for it. Poul Anderson explained the same thing in, well, slightly different terms back in 1989:

Unclefts with the same tale of firstbits but unlike tales of neitherbits are called samesteads. Thus, everyday sourstuff has eight neitherbits with its eight firstbits, but there are also kinds with five, six, seven, nine, ten, and eleven neitherbits. A samestead is known by the tale of both kernel motes, so that we have sourstuff-13, sourstuff-14, and so on, with sourstuff-16 being by far the most found. Having the same number of bernstonebits, the samesteads of a firststuff behave almost alike minglingly. They do show some unlikenesses, outstandingly among the heavier ones, and these can be worked to sunder samesteads from each other.

One other crazy thing about isotopes (or samesteads if you prefer). Not all neutrons have the same mass. The 13C atom’s mass 13.0034 daltons (Da), and the 12C atom’s is 12.0000 Da (by definition). Since those atoms differ only by a neutron, it would seem that a neutron in a carbon atom has a mass of 13.0034 – 12.0000 = 1.0034 Da. But hydrogen also has two isotopes: 2H with one neutron and one proton, and 1H with no neutrons at all, only a single proton. But a 2H atom’s mass is 2.0141 Da, and a 1H atom’s is 1.0078 Da. If you do the same subtraction, you’ll find that the mass of neutron in a 2H atom is 1.0063 Da. Those numbers are close but they are NOT the same. A neutron in a hydrogen atom is about 0.3% heavier than a carbon atom’s neutron.

The reason for the difference has to do with Einstein’s famous mass-energy relationship and the strong nuclear force and…well, let’s make it a subject for another time. For now, let’s end by saying that these seemingly slight differences in neutron mass from atoms of one element to another — called the mass defect — can be detected in modern mass spectrometers. And what of it, you say? Well, I’m glad you asked: that will be the subject of some upcoming posts. (Here’s a taste.)

CODATA values for physical constants in [R]

The R language (i) is widely used by physical, chemical, and biological researchers and (ii) comes with lots of interesting and easily loaded data sets. Unfortunately and surprisingly, a data set of commonly encountered physical constants is not one of them. If you’re writing R code to transform data based on physical equations, chances are these equations will involve physical constants such as Planck’s constant or the Newtonian gravitational constant.

Examples of data bundled in R include mtcars and other toy data that pop up in examples and tutorials everywhere. They can be loaded via commands like data(mtcars). It seems natural to have a data set for physical constants that could be loaded similarly, but as far as I know there isn’t one. And it isn’t just me that’s been wondering; the issue has come up on StackOverflow more than three years ago.

The best source for values of physical constants in NIST, whose Committee on Data for Science & Technology (CODATA) routinely aggregates data from experiments which measure a constant and adjusts all the aggregated values so the overall data set is self consistent. As of today it seems like the most recent update was from 2010. NIST data for the physical constants is browsable on their web page; a flat ASCII file is here.

Here’s some R code for transforming this ASCII table into an R data frame.

#parsing NIST CODATA data for physical constants 
require(stringr)	#simplifies regex usage syntax in R

#web page with data
link <- ''

allstr <- readLines('link')
codata.str   <- allstr[11:345]  #data starts at line 11

#at least 3 spaces separate columns
codata.str2  <- str_replace_all(codata.str, '[ ]{3,100}', '\t') 

#four cols in source table
codata.mat 	 <- str_split_fixed(codata.str2, '\t', n=4) 

#eliminate spaces separating every three decimal digits 
codata.mat[,c(2,3)] <- str_replace_all(codata.mat[,c(2,3)], " ", "") 

codata <- data.frame(quantity 	 = codata.mat[,1],
		     value	 = as.numeric(codata.mat[,2]),
		     uncertainty = as.numeric(codata.mat[,3]),
		     units 	 = codata.mat[,4]

And that’s it. Now it’s easy to search — in R — the data to find the constants you want, and assign constants that you will use to variable names, without having to manually type them.

> codata[str_detect(codata$quantity, 'electron mass'),]
                                  quantity        value uncertainty units
2       alpha particle-electron mass ratio 7.294300e+03     2.9e-06      
63            deuteron-electron mass ratio 3.670483e+03     1.5e-06      
89                           electron mass 9.109383e-31     4.0e-38    kg
90         electron mass energy equivalent 8.187105e-14     3.6e-21     J
91  electron mass energy equivalent in MeV 5.109989e-01     1.1e-08   MeV
92                      electron mass in u 5.485799e-04     2.2e-13     u
130             helion-electron mass ratio 5.495885e+03     5.0e-06      
195               muon-electron mass ratio 2.067683e+02     5.2e-06      
223            neutron-electron mass ratio 1.838684e+03     1.1e-06      
264             proton-electron mass ratio 1.836153e+03     7.5e-07      
310                tau-electron mass ratio 3.477150e+03     3.1e-01      
320             triton-electron mass ratio 5.496922e+03     5.0e-06 

> me.u < codata[str_detect(codata$quantity, 'electron mass in u'),'value']
> me.u
[1] 0.0005485799

I like the dataframe format because it’s what I (and I think many other R users) am used to, but there are other useful solutions at StackOverflow.

I wonder how to get the core R development team to include this listing or a similar one in base R for easy loading via a command like data(codata)

Raman spectroscopy is awesome and you should too


Sir C.V. Raman, the namesake of Raman spectroscopy. Image from Wikipedia.

Raman spectroscopy is awesome. But these days it is rapidly getting more awesome. Here are a just a few new reasons to love Raman spectroscopy:

  1. It measures the metabolism of single living bacterial cells (in conjunction with stable isotopes).
  2. In mere seconds, it quantifies essentially all the various gases in 27 nanoliters of human breath.
  3. It does video-speed label-free chemical imaging on microscopic scales.
  4. It tracks in real-time the yields and titers of chemical and bichemical reactions, even at small scales.

Continue reading

Open seminars: a new and good idea

One of the things I liked best about being an academic researcher was group meeting. Every week, a different student or scholar would present some fresh data from their own projects. And these meetings were casual and interactive: you could interrupt any time with questions.

In industry chances to keep up with fundamental discovery science outside of your own core area can be more limited. Folks (mainly job candidates) do visit the corporate world to give research talks, and of course industrial scientists still attend conferences, but those interactions — while invaluable — aren’t as informal. The work presented is always well-polished, and people usually shy away from long technical questions and discussion.

That isn’t the case with MicroSeminar: it’s a new(ish) online-only, publicly accessible research seminar in microbiology created by Jennifer Biddle and Cameron Trash. Once a month or so, people from all over the world log into a Google Plus hangout, or watch YouTube — live or when you get free time — as some of the new fresh hot research in environmental microbiology, microbial ecology, and biogeochemistry gets presented. The feel is informal, with lots of Q&A, and you don’t even have to leave home (or bed!). Here are some of the talks I’ve enjoyed so far.

And don’t just take my word for it. Here are some similar thoughts from Pat Schloss:

The cost of going to ISME [a conference] in Korea this summer? In the thousands. Cost of sitting with your laptop watching a seminar? Zilch. Jennifer is correct that this won’t kill conferences. Conferences have a huge social aspect and provide great opportunities for networking. But the science is frequently stale and pulled from the pages of last year’s AEM [a journal]. I think there’s great potential with this model to change how we disseminate information to our colleagues. Like I said, I think this is big, deserving of your attention and perhaps others will create parallel online seminar series that are either more specialized or more general.

And if microbiology isn’t your cup of tea? If you’re an academic in a different field? Like Pat says, maybe you should start another online seminar program like this one.

Learning to listen to the silent majority


Image from here.

Ninety-nine percent. That is the fraction of microbes that aren’t cultivatable by microbiologists. Throughout Earth’s ecosystems, they are there, living, breathing, metabolizing…but because microbiologists can’t grow them in their labs, figuring out what this “silent majority” of microbes do has been tough.

But new lots of new research is changing that. First, microbiologists have been cooking up (literally!) new ways to grow microbes. The old recipes for growth medium turned out to be toxic for many species of bacteria. Second, in many cases, it’s turned out that cultivation is possible, with sufficient time and attention to detail. (Breaking out the microscope is almost never a bad idea.)

Lastly and most impressively, a team has demonstrably slashed the fraction of “uncultivatable” microbes from 99% to something more like 50%…by cultivating the bacteria. Their approach to this impressive success relies on a microfluidic cultivation device they call iChip. These chips are made of hydrophobic plastic arrayed with holes about 1 μL in volume. The holes are open on both top and bottom. To use it, the chip is immersed in a suspension of microbes from environmental samples, so each hole fills up with (on average) about one cell from the sample. Each side of the chip is then sealed with permeable membrane filter. The filter has pore sizes big enough to let soluble nutrients in, but small enough to block microbial cells from leaving (0.03 μm filters seems to do the trick).

Next comes the key step. The sealed chip is returned to the environment from which the sample came for several weeks — the waiting game begins. In their latest study, the team looked at soil bacteria from a grassy field in Maine, so they buried their seeded chip back into Maine soil and left there for a whole month.

New colonies of microbes, unrelated to anything previously cultivated in a lab, grow in many of iChip’s wells. In 2010, the team found an uncultured variant of the marine bacterium Maribacter polysiphoniae that would grew on the chips, but only in proximity to another bacterium they isolated, Micrococcus luteus strain KLE1011. The were able to work out the reason for the co-dependence: M. polysiphonae needed iron-binding siderophores produced by M. polysiphonae to live. In the study published today by Nature, they found a new betaproteobacterium that they called Eleftheria terrae. This bacterium turned out to be tremendously exciting because it produced a new antibiotic, as detailed very well in an accompanying News & Views article.

The chip-based cultivation is clearly useful in understanding what that supposedly “uncultivable” 99% is doing. It’s fun to speculate as to why. It could simply be that microbiologists, a patient lot by most human standards, don’t usually wait long enough for colonies to grow. (One month is a long time!). In some cases, the chip’s holes are small enough so that neighboring cells can still influence growth, such as by supplying those key siderophores. Another possibility is that isolates survive their initial seeding in the chip because it’s an environment they’re accustomed to, but over time as the microcolony grows, they adjust — either through mutation or through somatic adaptation — to life in monoculture. Or it could something else entirely, and oh yes, the answer probably varies from species to species.

Regardless, I expect to be hearing many more reports on cultivating the many “uncultivable” microbes on Earth. As many have previously observed, if we listen correctly, we should indeed be able to hear the microbial silent majority, even as they grow in our labs.