Long-term Experiments

I recently read this Nature article, where is described what is probably one of the longest experiments ever to be conducted. A population of E. coli was kept for 20 years (!) in a nutrient solution (low on glucose), and samples were taken and deep-frozen after 2000, 5000, 10000, 15000, 20000 and 40000 generations. The authors sequenced the genome of the sample bacteria to investigate the rate of mutations.

Up to generation 20K, the number of mutations grew steadily to a total of 45. The adaptation to the environment, however, only increased strongly in the beginning. It was concluded that the most beneficial mutations were the first to occur. After generation 20K, a change in the mutT gene caused a rapid increase in the mutation rate to result in 653 mutation at generation 40K, but with a neutral signature, i.e. no further adaptation.

What I find most fascinating about this extreme long-term experiment is the confidence of the researchers that it would be possible to analyze the genes at a later point; this was not at all self-evident in the late ’80s! In addition, some work had to be done each day, for twenty years. What if the power had failed for a week or so? Of course, this unique opportunity to watch evolution as it happens is very intriguing.

An experiment that took even longer was awarded this year’s Ig Nobel Prize in medicine: Donald L. Unger of Thousand Oaks, CA, cracked the knuckles of his left hand, but not his right hand, every day for 50 years to see if knuckle-cracking leads to arthritis. After this time, both hands were fine, so he concluded: “While a larger group would be necessary to confirm this result, this preliminary investigation suggests a lack of correlation between knuckle cracking and the development of arthritis of the fingers.” Apparently, the experiment must be repeated.

Subscribe to the comments for this post

Joking in a Nature journal?

Was reading this earlier this evening on the hobby science forum Sciencemadness.org –it’s amusing.

Read the letter carefully.

Well, let’s hope this is a joke.

Subscribe to the comments for this post

Condensed Print Format

My boss has pointed out this piece of news covered by C&EN. Apparently, starting from July, all ACS journals will be printed in a “rotated and condensed” format, that is two pages on one printed page in landscape format. This is an effort to reduce printing and distribution costs.

In my opinion, this change is just one further step towards purely electronic journals that are not printed at all. I think this will deeply affect the way we present our data and how we look at formatting. Preparing a manuscript in a way meant for printing is different from one which will never appear in print. Some may welcome this change because it saves paper, others will probably miss the possibility of flicking through a new issue of JACS. Although I rarely go to the library to pick up a printed journal, I admit to reading printouts very often (see this post).

Update: Apparently, in 2010 the print versions will stop completely, with the exception of JACS, Acc. Chem. Res. and Chem. Rev. See also Nature News.

Subscribe to the comments for this post

Science Podcasts

I got a new iPod Touch from Sigma-Aldrich for Christmas this year (no, really).  Probably the best part is the wi-fi capabilities.  I tried keeping up with podcasts before on other iPods, but it was too much work to plug the iPod into the computer, transfer over the new episodes, and repeat.  Now I can grab the new episodes directly from the iTouch, and it’s awesome.  (btw, does anyone know how to subscribe to the podcasts so the iPod will update itself with new episodes automatically?)

Anyway, I’ve been surveying the chemistry podcasts over the past few weeks.  They’re great to listen to while running a column or doing other tasks that don’t require a lot of mental energy.  Most of the major journals have a podcast, as do most news outlets and some other random sites.  I’ll tell you about some of the ones I liked below the jump if you’re interested in giving some of them a listen.

Read more »

Subscribe to the comments for this post

Hoveyda’s Asymmetric Mo-Catalyzed Metathesis

I recently had the pleasure of listening to an excellent talk by Amir Hoveyda about his chiral-at-metal Mo catalysts for asymmetric metathesis. This kind of catalyst is based on the Schrock-type molybdenum catalysts. Most asymmetric catalysts nowadays employ bidentate ligands such as BINOL-type ligands that carry the stereochemical information. By chelating the metal centre, the fluxionality¹ is reduced, therefore ensuring a well-defined geometrical arrangement of the catalyst-substrate complex and good stereoselectivity. In addition, the loss of stereochemical integrity of the catalyst is suppressed. Hoveyda has used such complexes to achieve the formation of P-stereogenic phosphinates and phosphinoxides (where the phosphorus is a chiral centre) by desymmetrizing RCM of achiral precursors (ACIE: 10.1002/anie.200805066).

There is, however, a downside to reduced fluxionality. As Hoveyda points out, the Mo complex has to undergo a series of geometrical rearrangements during a catalytic cycle. If the complex is too rigid, these rearrangements are hindered. In other words, you may get high e.e.’s, but the catalytic activity will be low. This is where chiral-at-metal catalysts have a clear advantage: they only require monodentate ligands, which makes the complex more able to rearrange, and the reaction will be much faster.

The excellent stereocontrol is due to electronic effects rather than pure sterics. One of the ligands has to be an acceptor (the BINOL-type ligand) that ensures sufficient Lewis acidity of the metal centre. A donor ligand (the pyrrole) is also required because it distorts the complex geometrically in a way that facilitates the coordination of the alkene substrate. The use of this kind of asymmetric RCM is demonstrated in a total synthesis of (+)-quebrachamine, where they get a yield of 84% and an excellent 96% e.e. (Nature: 10.1038/nature07594).

I have only touched some of the most important points made in these papers. They are definitely worth reading!

¹ I wasn’t familiar with the term “fluxionality”. Apparently, it is often used in organometallic chemistry to indicate the possibility of interchanging between equivalent conformational arrangements. As a simple example, Wikipedia mentions the interchange of the two Me groups in dimethylformamide. A monodentate ligand in a metal complex is fluxional in the sense that it can rotate around the metal-ligand bond.

Subscribe to the comments for this post

Linear-Motor from Carbon Nanotubes

A recent paper this week by Somada et al. regarding making a linear-motor from carbon nanotubes piqued my interest.^[NanoLett] The general design idea is to encapsulate a piece of carbon nanotube material within a larger carbon nanotube. If done correctly you can end up with a configuration as shown below.

The cargo, in yellow, transverses the carbon nanotube and rests in either position A or position B. An abridged summary of their observations is as follows: 1) From observing the system for 170 s the cargo traveled back-and-forth seven times; 2) The cargo was never filmed in between positions A and B, indicating the movement was less than the frame rate (0.5 s). From this information I can construct a likely energy landscape for this system.

Mitch’s hypothetical potential energy map for the linear-motor.

The diagram replicates the observation that the cargo at room temperature will be trapped at either position A or B. It also explains why it’s never seen between A or B, as there is no energy minimum for it to rest in. Lets assume every ~20 s there is randomly enough thermal energy to kick the system over the barrier, and that this accessible energy exists for less than 0.5 s. Then you would expect the cargo to be able to move either to A or B, and to do it faster than the shutter speed.

This is an interesting system for analysis, but it’s not a motor. Or conversely, it is as much a motor as ethane is a useful rotor. Just because thermal energy provides the means for things to happen it doesn’t mean it generates usable work. There is no way to construct a usable motor or any device from this system, but it’s a first step in that direction. I suspect if the authors raised the temperature they would see the cargo undergoes random walk motion. Thermal energy yields a random linear-motor.

Link to article: A Molecular Linear Motor Consisting of Carbon Nanotubes

Update 1: Tim Reid also covered it at Nature Chemistry — Nanotube motors: Sliding and spinning

Mitch

Subscribe to the comments for this post

A Word on Research Misconduct

Dig out your dictionary and look up the word “hyperbole” (I know, it might be a while since you’ve last had English class)—exaggerated statement or claims not intended to be taken seriously.  I tend to hyperbolize a bit when I replay an incident that happened at the bar or in class, which I attribute to the fact that I’m a terrible storyteller.  I think we all do it to a certain extent.  I know I’ve once said something to the effect of, “It was the greatest movie, ever…in the history of humans.”  A hyperbole at its finest.

While most common vernacular is riddled with hyperboles, I’d argue that the majority of intellectual study makes an effort to stay away from gross exaggerations (with history being the exception).  In particular, science is the observation and study of the physical world, and it leaves no room for hyperboles.  Just facts.  For example, if you mix an aqueous solution of silver nitrate with an aqueous solution of sodium chloride it is a fact that a precipitate will form.  There are no equivocations about scientific facts.  Though, science sometimes falls short when making assumptions that connect two or more facts into one coherent theory or proposal.  Still, these assumptions, en route to a new theory, are usually reasonable if not simplistic (i.e. Occam’s razor).

What about bad data?  Of course, there are ways to make our raw data more “natural” without exaggerating.  In the event that we have to plot data points, for example, as scientists we can exclude data that “doesn’t belong.”  We call these anomalies “outliers” and there is statistical rationale as to why a stray might be “bounced” from the data set without any bias to the result.  But even in these cases, the data point is often so far away from the others that including it might be a detriment to a fact about Mother Nature.

What irritates me to no extent is a term I refer to as “hyperbolized research.”  We have all seen these situations before: yields that are bumped a good 5 to 10 to 50%, data that is fit just right, patent procedures that are not reproducible.  Why are these practices tolerated?  Contemporary science is themed “publish or perish,” which essentially means that if you are not producing enough results (nevermind quality) you will soon be unemployed.  I recall hearing stories about early 20^th century scientists who studied science without the proverbial gun to their respective heads and still made great findings.  A lot of these experiments were groundbreaking, marvelous and truly beautiful.

It’s no surprise that this issue of “publish or perish” rears its ugly head in science.  Society is incredibly fast-paced, and science is certainly trying to keep up.  But, it’s really hard to do so with a tiny, bankrupt research group (where most if not all members are teaching) versus a behemoth firm with hundreds of years of experience and millions of dollars of materials to use. 

So, what do groups do to keep pace (or at least appease the boss)?  “How did you do with that reaction you couldn’t get to work last week?”  “Um…I got 98% yield with 95% ee.”  “Great, let’s write up a manuscript and submit to JACS.”  I’ve heard stories of “big name” research groups who’s members purposely inflate their yields to keep “the man” happy.  In these cases, researchers keep two sets of lab notebooks: the real one (usually under lock and key with the actual experimental results) and the boss’ one (usually kept in the open, so the boss can see how his researcher got a 90% yield on chemistry that is next to impossible to reproduce).  The bottom line is that papers get published, lectures are given and proposals are funded—criminality is rewarded.  How is this right?  Furthermore, how is it fair to another researcher who needs to repeat the results?

Have we not learned anything from the Bell Lab incident?  For those not familiar, Hendrik Schön was a groundbreaking physicist working for Bell Labs in the late 1990’s.  He was purportedly on par to win a Nobel Prize with his creation of an “organic molecular transistor.”  The papers describing this work were met with criticism in the scientific community and at some point (c. 2001), Bell Labs launched an internal investigatory committee to examine Schön’s work.  Their final report ultimately alleged 24 accounts of misconduct that were essentially fit into three categories: “Substitution of data,” “unrealistic precision of data,” “Results that contradict known physics.”  In the end, he was ultimately stripped of his doctoral degree.  But think about the repercussions of not investigating Schön’s findings.  Had Schön’s work not been policed, potentially millions of dollars would’ve been invested into falsified research.  While I’m aware that Bell Labs was recently closed, without insinuating anything, it makes me wonder if this Schön incident had any weight in the lab’s termination.

Rex Dalton covered the aftermath of this incident along with several other examples of research misconduct (Nature 2002, 420, 728-729).  He ultimately offered up the following observation:

“Science may be self-correcting, but sometimes it is a painfully slow process.”

Perhaps he’s right.  Sure, several papers are going to be questioned in the future.  And of those papers, a few might be blatant lies.  How much time is it going to take to correct these mistakes?  According to Corey: “Occasionally, blatantly wrong science is published, and to the credit of synthetic chemistry, the corrections usually come quickly and cleanly.”  Case in point?  The hexacyclinol incident that was excellently covered by C&EN and by a couple of fellow bloggers: Derek Lowe and Paul Docherty.  In this case, there was a rapid turn around (possibly due to public interest).  However, this case might be the exception.  It could be years before a questionable project is proven incorrect.

I know…you want me to provide a solution.  Maybe there isn’t an immediate, reasonable answer.  But, alas, here’s what I’ve uncovered: there are a few wonderful articles in J. Chem. Ed. about scientific misconduct, which both hover around the LBNL and Bell Lab incidents (see: J. Chem. Ed. 2002, 79, 1391; ibid. 2005, 82, 1521).  The authors’ messages (albeit bluntly or implied) were that ethics and empathy should be at the forefront in the early years of scientific training.  Some people cannot discern between right and wrong and teachers should do their jobs by teaching students about the rights and responsibilities of being a scientist.  While I did not receive formal training on scientific misconduct, I was given a lambasting for bordering on plagiarism my freshman year of college.   I learned my lesson early—you and your lab partner need to keep separate lab notebooks.  Perhaps this experience has formed me into the scientist that I am today (I’m anal-retentive about my lab notebook). 

I guess there is a remaining question still looming.  What sparked this rant about “doing the right thing”?  I’ve been repeating experiments for the past couple of months that were reported to be exceptionally clean (requiring no chromatography) and high yielding.  Most of these reactions have tanked—miserably—even with exceptional preparation and precision.  So, I’m painstakingly re-optimizing experimental procedures so someone else doesn’t have to.  It’s taking a while—much longer than it reasonably should.  But, hey, “sometimes (correcting science) is a painfully slow process.” 

Subscribe to the comments for this post

Scientific Misconduct

This nature article discusses the results of a survey about scientific misconduct, while an editorial makes some comments.

Quote: “The 2,212 researchers we surveyed observed 201 instances of likely misconduct over a threeyear period. That’s 3 incidents per 100 researchers per year. A conservative extrapolation from our findings to all DHHS-funded researchers predicts that more than 2,300 observations of potential misconduct are made every year.” Almost 9% of the respondents had witnessed some sort of misconduct, and 37% of those incidents went unreported.

The authors conclude that, besides protecting the whistleblowers better, it is necessary “to create a zero-tolerance culture”. The editor, however, holds the opinion that one also needs to take a look at “the environment that has allowed misconduct to flourish”. In his opinion, there should be the possibility of finding a solution without ruining the career of a scientist, especially in mild cases.

I tend to follow the editor’s reasoning. In my opinion, the zero-tolerance culture already exists to a certain extent, because a scientist convicted of, e.g. faking data, can forget about his career. But the result of such a policy is clear: no-one wants to blow the whistle on a colleague, because they don’t want to end somebody else’s career and because they will make themselves very unpopular. The real problem is the way misconduct is treated at the moment: we want to identify the guilty scientist, and punish him/her.

While this makes sense for the worst cases of fraud, in milder cases one should try and ask the question *why* the misdeed was done. Take, for example, the way hospitals treat mistakes nowadays: they try to find out how it could happen, and how it can be avoided in the future. This is very sensible, because it treats the problem in a proactive way: instead of reacting to the incident by punishing somebody, future incidents are reduced by tackling the things that cause them in the first place.

If there is a lot of pressure to produce as much data as possible in a research group, it is tempting to cut a corner once in a while. Can this not partly be considered the prof’s fault? In a similar way, one should address the working atmosphere in the group in question. The problem with the academic system is that there is no informal institution to turn to, besides your boss, if you are to witness a case of scientific misconduct. So we fall back to the old issue: the only person you can contact in case of problems has all the power over you.

At the University of Toronto, a “Graduate Student Oath”, similar to the Hippocratic Oath, has been tried as a means to strengthen scientific ethics (Science). Although this is an interesting idea, I doubt it will change the behaviour of people very much.

Subscribe to the comments for this post

Social Bookmarking for Scientists: HotCites

HotCites – It’s What’s Hot in the Literature

A recent website I’ve been developing exploits the collective intelligence of scientists to determine what articles in the literature are attracting the attention of scientists/technicians and the like. The website is HotCites (www.hotcites.com). It parses through new submissions to Nature’s Connotea database and spits out what users (scientists) consider important enough to add to their libraries. I could talk forever about how great Connotea is, the most important selling points for me are the following: Import/Export your database to endnote in less than a minute, see who else has bookmarked an interesting paper, and then view their literature databases, search for papers that others have tagged (ie aromatic, metabolomic), and access to your literature database from any internet connection.

Keeping a database of people’s libraries is all well and good, but there is untapped information that can be extracted. Both CiteULike and Connotea seem to have the databasing side down, but an engaging user environment seems lacking from my tedious perusing. Creating an expanding a community of users is always the most time mitigating part of any successful Web 2.0 website.

So lets move forward and make a more cheerful website, where we’ve already distilled some information from a database and present it back to the end users. A thumb of the front page of HotCites is shown below:

The website immediately has distilled what Connotea users have bookmarked the most frequently for the past week. You can also click to see what the top bookmarked papers are for the day. The database currently only goes back 17 days or so, but as time pases the database will enlarge, and I’ll be making options to go further back. Be mindful that the url to see the past “2 weeks” is just http://www.hotcites.com/index.php?&days=14 , so you can edit the number of days to go back as far as you like. But as I mentioned, the database only goes back 17 days or so.

Getting HotCites fully integrated with Connotea will be mission #1. Hopefully the appealing(in my eyes) nature of the website design we’ll encourage more scientists to use Connotea. An advantage of HotCites is it’ll only show Connotea bookmarks that have resolved DOIs, in order to ensure only literature gets posted. But, editorials also have DOIs, and they will show up too. Expansion to incorporate CiteULike may be possible down the road, but I’m unsure how liberal they are with their database. I hope you like the website, now go sign up for Connotea!

Mitch

Subscribe to the comments for this post

Mitch @ NatureJobs

Nature Jobs did a nice piece on Social Networking for scientists recently, I even get two nice mentions in it. The article is titled “The new networking nexus” and can be found here:

http://dx.doi.org/10.1038/nj7181-1024a

Mitch

Subscribe to the comments for this post