Category Archives: Humor
The format of a research paper hasn’t changed much throughout history, despite the enormous changes in platforms available for their consumption and the near extinction of the library issue. Convenient electronic files such as PDFs still resemble printed-and-bound-in-issues papers in their layout instead of exploiting the seemingly endless capabilities of the electronic format.
For instance, why do we still need to have page numbers? a DOI is a full, traceable and unique identification for each work and there are so many nowadays that publishers have to pour them out as e-first, ASAPs, and just accepted before having them assigned page numbers, a process which is still a concern for some researchers (and even for some of the organizations funding them or evaluating their performance). Numbers for Issues, Volumes and Pages are library indexes needed to sort and retrieve information from physical journals but in the e-realm where one can browse all issues online, perform a search and download the results these indexes are hardly of any use, only the year is helpful in establishing a chronological order to the development of ideas. This brings me to the next issue (no pun intended): If bound-issues are no longer a thing then neither should be covers. Being selected for a cover is a huge honor, it means the editorial staff think your work stands out from the published works in the same period; but nowadays is an honor that comes to a price, sometimes a high price. With the existence of covers, back-covers, inner-covers and inner-back-covers and whatnot at USD$1,500 a piece, the honor gets a bit diluted. Advertisers know this and now they place their ads as banners, pop-ups and other online digital formats instead of -to some extent- paying for placing ads in the pages of the journals.
I recently posted a quick informal poll on Twitter about the scientific reading habits of chemists and I confirmed what I expected: only one in five still prefers to mostly read papers on actual paper*, the rest rely on an electronic version such as HTML full text or the most popular PDF on a suitable reader.
— Joaquin Barroso (@joaquinbarroso) June 3, 2019
What came as a surprise for me was that in the follow up poll, Reference Manager programs such as Mendeley, Zotero, EndNote or ReadCube are only preferred by 15% while 80% prefer the PDF reader (I’m guessing Acrobat Reader might be the most popular.) A minority seems to prefer the HTML full text version, which I think is the richest but hardly customizable for note taking, sharing, or, uhm hoarding.
A follow up on the previous poll. Dear #ChemTweeps, if you mostly read papers in electronic format what is your preferred platform?
— Joaquin Barroso (@joaquinbarroso) June 10, 2019
I’m a Mendeley user because I like the integration between users, its portability between platforms and the synchronization features but if I were to move to another reference manager software it would be ReadCube. I like taking notes, highlighting text, and adding summaries and ideas onto the file but above all I like the fact that I can conduct searches in the myriad of PDF files I’ve acumulated over the years. During my PhD studies I had piles of (physical) paper and folders with PDF files that sometimes were easier to print than to sort and organize (I even had a spreadsheet with the literature read-a nightmarish project in itself!)
So, here is my wish list for what I want e-papers in the 21st century to do. Some features are somewhat available in some journals and some can be achieved within the PDF itself others would require a new format or a new platform to be carried out. Please comment what other features would you like to have in papers.
- Say goodbye to the two columns format. I’m zooming to a single column anyway.
- Pop-up charts/plots/schemes/figures. Let me take a look at any graphical object by hovering (or 3D touching in iOS, whatever) on the “see Figure X” legend instead of having to move back and forth to check it, specially when the legend is “see figure SX” and I have to go to the Supporting Information file/section.
- Pop-up References. Currently some PDFs let you jump to the References section when you click on one but you can’t jump back but scroll and find the point where you left.
- Interactive objects. Structures, whether from X-ray diffraction experiments or calculations could be deposited as raw coordinates files for people to play with and most importantly to download** and work with. This would increase the hosting journals need to devote to each work so I’m not holding my breath.
- Audio output. This one should be trickier, but far most helpful. I commute long hours so having papers being read out loud would be a huge time-saver, but it has to be smart. Currently I make Siri read papers by opening them in the Mendeley app, then “select all“, “voice“, but when it hits a formula, or a set of equations the flow is lost (instead of reading water as ‘H-Two-O‘, it reads ‘H-subscript Two-O‘; try having the formula of a perovskite be read)
- A compiler that outputs the ‘traditional version‘ for printing. Sure, why not.
I realize this post may come out as shallow in view of the Plan-S or FAIR initiatives, sorry for that but comfort is not incompatible with accessibility.
What other features do you think research papers should have by now?
* It is true that our attention -and more importantly- our retention of information is not the same when we read on paper than on a screen. Recently there was an interview on this matter on Science Friday.
** I absolutely hate having a Supporting Information section with long PDF lists of coordinates to copy-paste and fix into a new input file. OpenBabel, people!
To chem or not -quite- too chem, that is the ChemNobel question:
Whether ’tis Nobeler in the mind to suffer
The curly arrows of organic fortune
Or to take rays against a sea of crystals
And by diffracting end them.
Me (With sincere apologies to WS)
Every year, in late September -like most chemists- I try to guess who will become the next Nobel Laureate in Chemistry. Also, every year, in early October -like most chemists- I participate in the awkward and pointless discussion of whether the prize was actually awarded to chemistry or not. Indeed, the Nobel prize for chemistry commonly stirs a conversation of whether the accomplishments being recognized lie within the realm of chemistry or biology whenever biochemistry shows its head, however shyly; but the task of dividing chemistry into sub-disciplines raises an even deeper question about the current validity of dividing science into broad branches in the first place and then further into narrower sub-disciplines.
I made a very lazy histogram of all the 178 Laureates since 1904 to 2017 based on subjective and personal categories (figure 1), and the creation of those categories was in itself an exercise in science contemplation. My criteria for some of the tough ones was the following: For instance, if it dealt with phenomena of atomic or sub-molecular properties (Rutherford 1908, Hahn 1944, Zewail 1999) then I placed it in the Chemical Physics category but if it dealt with an ensemble of molecules (Arrhenius 1903, Langmuir 1932, Molina 1995) then Physical Chemistry was chosen. Some achievements were about generating an analysis technique which then became essential to the development of chemistry or any of its branches but not for a chemical process per se, those I placed into the Analytical Chemistry box, like last year’s 2017 prize for electron cryo-microscopy (Dubochet, Frank, Henerson) or like 1923 prize to Fritz Pregl for “the invention of the method of microanalysis of organic substances” for which the then head of the Swedish Academy of Sciences, O. Hammarsten, pointed out that the prize was awarded not for a discovery but for modifying existing methods (which sounds a lot like a chemistry disclaimer to me). One of the things I learnt from this exercise is that subdividing chemistry became harder as the time moved forward which is a natural consequence of a more complex multi- and interdisciplinary environment that impacts more than one field. Take for instance the 2014 (Super Resolved Fluorescence Microscopy) and 2017 (Cryo-Electron Microscopy) prizes; out of the six laureates, only William Moerner has a chemistry related background a fact that was probably spotted by Milhouse Van Houten (vide infra).
Some of the ones that gave me the harder time: 1980, Gilbert and Sanger are doing structural chemistry by means of developing analytical techniques but their work on sequencing is highly influential in biochemistry that they went to the latter box; The same problem arose with Klug (1982) and the Mullis-Smith duo (1993). In 1987, the Nobel citation for Supramolecular Chemistry (Lehn-Cram-Pedersen) reads “for their development and use of molecules with structure-specific interactions of high selectivity.”, but I asked myself, are these non-covalent-bond-forming reactions still considered chemical reactions? I want to say yes, so placed the Lehn-Cram-Pedersen trio in the Synthesis category. For the 1975 prize I was split so I split the prizes and thus Prelog (stereochemistry of molecules) went into the Synthesis category (although I was thinking in terms of organic chemistry synthesis) and Cornforth (stereochemical control of enzymatic reactions) went into biochem. So, long story short, chemistry’s impact in biology has always had a preponderant position for the selection of the Nobel Prize in Chemistry, although if we fuse the Synthesis and Inorganic Chemistry columns we get a fairly even number of synthesis v biochemistry prizes.
Hard as it may be to fit a Laureate into a category, trying to predict the winners and even bet on it adds a lot of fun to the science being recognized. Hey! even The Simpsons did it with a pretty good record as shown below. Just last week, there was a very interesting and amusing ACS Webinar where the panelist shared their insights on the nomination and selection process inside the Swedish Academy; some of their picks were: Christopher Walsh (antibiotics); Karl Deisseroth (optogenetics); Horwich and Hartl (chaperon proteins); Robert Bergman (C-H activation); and John Goodenough (Li-ion batteries). Arguably, the first three of those five could fit the biochem profile. From those picks the feel-good prize and my personal favorite is John Goodenough not only because Li-ion batteries have shaped the modern world but because Prof. Goodenough is 96 years old and still very actively working in his lab at UT-Austin (Texas, US) #WeAreAllGoodEnough. Another personal favorite of mine is Omar Yaghi not only for the development of Metal-Organic-Frameworks (MOFs) but for a personal interaction we had twenty years ago that maybe one day I’ll recount here but for now I’ll just state the obvious: MOFs have shown a great potential for applications in various fields of chemistry and engineering but perhaps they should first become highly commercial for Yaghi to get the Nobel Prize.
Some curiosities and useless trivia: Fred Sanger is the only person to have been awarded the Nobel Prize in Chemistry twice. Marie Curie is the only person to have been awarded two Nobel Prizes in different scientific categories (Physics and Chemistry) and Linus Pauling was awarded two distinct Nobel Prizes (Chemistry and Peace). Hence, three out of the four persons ever to have been awarded two Nobel Prizes did it at least once in chemistry – the fourth is John Bardeen two times recipient of the Nobel Prize in Physics.
Of course the first thing I’ll do next Wednesday right after waking up is checking who got the Nobel Prize in Chemistry 2018 and most likely the second thing will be going to my Twitter feed and react to it, hopefully the third will be to blog about it.
The announcement is only two days away, who is your favorite?
As we were hanging out recently, the idea came to us at the lab to create memes in order to summarize our work. We should be writing articles but hey, we needed the break, and so we shared them with each other in our last group meeting along with a good laugh. Here are some of the funniest ones.
Having doughnuts during our weekly meetings has proven a huge success in itself:
Finding transition states for organic chemical reactions can be a bit frustrating at times:
Good old photosynthesis sparked a few realizations too:
We’re dealing with docking calculations for a massive number of molecules. This has sparked a few inside jokes too:
A conversation about heterocyclic nomenclature that sparked this other post:
Try your own and share. Thanks for reading.
Communication of scientific findings is an essential skill for any scientist, yet it’s one of those things some students are reluctant to do partially because of the infamous blank page scare. Once they are confronted to writing their thesis or papers they make some common mistakes like for instance not thinking who their audience is or not adhering to the main points. One of the the highest form of communication, believe it or not, is gossip, because gossip goes straight to the point, is juicy (i.e. interesting) and seldom needs contextualization i.e. you deliver it just to the right audience (that’s why gossiping about friends to your relatives is almost never fun) and you do it at the right time (that’s the difference between gossips and anecdotes). Therefore, I tell my students to write as if they were gossiping; treat your research in a good narrative way, because a poor narrative can make your results be overlooked.
I’ve read too many theses in which conclusions are about how well the methods work, and unless your thesis has to do with developing a new method, that is a terrible mistake. Methods work well, that is why they are established methods.
Take the following example for a piece of gossip: Say you are in a committed monogamous relationship and you have the feeling your significant other is cheating on you. This is your hypothesis. This hypothesis is supported by their strange behavior, that would be the evidence supporting your hypothesis; but be careful because there could also be anecdotal evidence which isn’t significant to your own as in the spouse of a friend had this behavior when cheating ergo mine is cheating too. The use of anecdotal evidence to support a hypothesis should be avoided like the plague. Then, you need an experimental setup to prove, or even better disprove, your hypothesis. To that end you could hack into your better half’s email, have them followed either by yourself or a third party, confronting their friends, snooping their phone, just basically about anything that might give you some information. This is the core of your research: your data. But data is meaningless without a conclusion, some people think data should speak for itself and let each reader come up with their own conclusions so they don’t get biased by your own vision and while there is some truth to that, your data makes sense in a context that you helped develop so providing your own conclusions is needed or we aren’t scientists but stamp collectors.
This is when most students make a terrible mistake because here is where gossip skills come in handy: When asked by friends (peers) what was it that you found out, most students will try to convince them that they knew the best algorithms for hacking a phone or that they were super conspicuous when following their partners or even how important was the new method for installing a third party app on their phones to have a text message sent every time their phone when outside a certain area, and yeah, by the way, I found them in bed together. Ultimately their question is left unanswered and the true conclusion lies buried in a lengthy boring description of the work performed; remember, you performed all that work to reach an ultimate goal not just for the sake of performing it.
Writers say that every sentence in a book should either move the story forward or show character; in the same way, every section of your scientific written piece should help make the point of your research, keep the why and the what distinct from the how, and don’t be afraid about treating your research as the best piece of gossip you’ve had in years because if you are a science student it is.
In a nutshell, computational chemistry models are about depicting, reproducing and predicting the electronic-based molecular reality. I had this conversation with my students last week and at some point I drew a parallel between them and art in terms of how such reality is approached.
Semi empirical methods
Prehistoric wall paintings depict a coarse aspect of reality without any detail but nevertheless we can draw some conclusions from the images. In the most sophisticated of these images, the cave paintings in Altamira, we can discern a bison, or could it be a bull? but definitely not a giraffe nor a whale, most in the same way Hückel´s method provides an ad hoc picture of π electron density without any regard of the σ portion of the electron density or the conformational possibilities (s-cis and s-trans 1,3-butadiene have the same Hückel description).
More sophisticated semi-empirical Hamiltonians like PM3 or PM6 have better parametrizations and hence yield better results. We are still replacing a lot of information for experimental or adjusted parameters but we still cannot truly adopt it as truthful. Take this pre-medieval painting of one of the first Kings of England, Aelred the Unready. It is, by today standards, a good children´s drawing and not a royal portrait, we now see more detail and can discern many more features yielding a better description of a human figure than those found in Altamira or Egypt.
HF is the simplest of ab initio methods, meaning that no experimental results or adjustable parameters are introduced. Even more so, from the HF equations for a multi-electron system that complies with Pauli’s exclusion principle the exchange operator arises as a new quantum feature of matter with no classical analogue. Still, there are some shortcomings. Correlation energy is disregarded and most results vary according to the basis set employed. Take the impressionist movement, specially in France: In Monet´s Lady with Umbrella we have a more complicated composition, we observe many more features and although we have a better description of color composition some details, like her face, remain obscure. The impressionists are characterized by their broad strokes, the thicker the strokes the harder it is to observe details similar to what happens in HF when we change from a small to a large basis set, respectively.
CI (Configurations Interaction)
Extension of HF to a multi-reference method yields better results. In CI we take the original guess wavefunction -as expressed through a Slater Determinant- and extend it with one or many more wavefunctions; thus a linear combination of Slater Determinants gives rise to a broader description of the ground state because other electronic configurations are involved to include more details like the ionic and covalent pictures (configurations). The more terms we include the more real the results feel. If we take classical figurative paintings we have a similar result; most of these paintings are constituted of many elements and the more realistically each element is captured the more real the whole composition looks even if some are just merely indicated.
CCSD(T) full-CI, CASPT2
In Edwards Much’s the scream, we might think we have lost some information again and went back to impressionism but we know this is actually an expressionist painting; we can now not only observe details of the figurative portion of the image but Munch has captured his subject´s fear in the form of distorsions on the subjective reality. In this way, CCSD(T), full-CI and CASPT2 methods provide a description of the ground as well as the excited states which -in experimental reality- are only accessed through a perturbation of the elecron density by electromagnetic radiation. Something resembling radiation has perturbated the subject in The Scream rendering him frightened and wondering how to return to his ground state or if such thing will be even possible.
Density Functional Methods
At least due to its widespread use, DFT has risen as the preferred method. One of the reasons behind its success is the reduced computing time when compared to previous ab initio methods. So DFT is pretty much like photography, in which reality is captured in full but only apparently after selecting a given lens, an exposition, a filter, shutter speed and the occasional Photoshop for correcting issues such as aliasing. In photography, as in DFT, all details concerning the procedure or method for capturing an uncanny reproduction of reality must be stated in every case for reproduction purposes.
Now, in the end it all comes down to Magritte’s Pipe. Ceci n’est pas une pipe -or, ‘this is not a pipe’- reminds us that painting as with modeling we don’t get reality but rather a depiction of it. In this famous painting we look at an image that in our heads resembles that of a pipe but we cannot grab it, fill it with tobacco and smoke it.
The image above is a digital file, which translated becomes a scaled reproduction of an image painted by Magritte in which we see the 2D projection of the image of an object that reminds us of a pipe. In fact, the real name of this work is The Treachery of Images, definitely quite an epistemology problem on perception and knowledge but before I get too metaphysical I should finish this post.
Can you find where cubism or surrealism should be placed? with MPn methods, perhaps?
Pauli’s Exclusion Principle is a paramount concept in Quantum Mechanics which has implications from statistical mechanics to quantum chemistry, consequently, there are many different statements to summarize it depending on the forum. I occasionally joke with my students about how we learnt it in kindergarten an how we state it now at the end of our computational chemistry course.
So, are you a toddler or high up there with W. Pauli predicting the existence of sub-atomic particles at CERN? Which statement of Pauli’s Exclusion Principle sounds more familiar to you?
LOL just feeling a little humorous this morning!