Slashdot Mirror
The Most Highly Cited Scientific Papers of All Time
bmahersciwriter writes Citation is the common way that scientists nod to the important and foundational work that preceded their own and the number of times a particular paper is cited is often used as a rough measure of its impact. So what are the most highly cited papers in the past century plus of scientific research? Is it the determination of DNA's structure? The identification of rapid expansion in the Universe? No. The top 100 most cited papers are actually a motley crew of methods, data resources and software tools that through usability, practicality and a little bit of luck have propelled them to the top of an enormous corpus of scientific literature.
The biologist journal editor Ann Körner distilled her experience into the handbook Guide to Publishing a Scientific Paper , which I read a few months back. To warn against overciting, she notes how many young researchers today are likely to cite the original 1950s Crick and Watson paper, even though DNA is familair enough to treat as a given. Is it? I would have assumed journals would let you err on the side of caution and simply remove your citation if it were unnecessary, but apparently citing too much can block approval.
This is what Newton meant when he talked about standing on the shoulders of giants. These methods, algorithms, computer programmes, techniques etc. enable all the research you hear of. The structure of DNA would never have been solved without all the preceding work on x-ray crystallography, for instance. This is truly a case of credit where credit's due and not something surprising...
Would printing the 1,5 meter stack of most influential papers be a reasonable alternative to printing the 109.91 GB Survival Library http://tech.slashdot.org/story/14/10/28/0531223/a-library-for-survival-knowledge ?
Alan Turing's most widely cited paper is in biology.
http://en.wikipedia.org/wiki/The_Chemical_Basis_of_Morphogenesis
The summary reads like it was written by Buzzfeed. It's also a top 100 list. Hovering over the link, it appears to be nature.com, but I'm going to err on the side of caution and not click on it.
1) PDF version http://devbio.wustl.edu/InfoSo...
2) Commentary, 2004: http://www.jbc.org/content/280...
3) Wikipedia: http://en.wikipedia.org/wiki/L...
"The Lowry protein assay is a biochemical assay for determining the total level of protein in a solution. The total protein concentration is exhibited by a color change of the sample solution in proportion to protein concentration, which can then be measured using colorimetric techniques. It is named for the biochemist Oliver H. Lowry who developed the reagent in the 1940s. His 1951 paper describing the technique is the most-highly cited paper ever in the scientific literature, cited over 200,000 times."
The method combines the reactions of copper ions with the peptide bonds under alkaline conditions (the Biuret test) with the oxidation of aromatic protein residues. The Lowry method is best used with protein concentrations of 0.01–1.0 mg/mL and is based on the reaction of Cu+, produced by the oxidation of peptide bonds, with Folin–Ciocalteu reagent (a mixture of phosphotungstic acid and phosphomolybdic acid in the Folin–Ciocalteu reaction). The reaction mechanism is not well understood, but involves reduction of the Folin–Ciocalteu reagent and oxidation of aromatic residues (mainly tryptophan, also tyrosine). Experiments have shown that cysteine is also reactive to the reagent. Therefore, cysteine residues in protein probably also contribute to the absorbance seen in the Lowry Assay. [3] The concentration of the reduced Folin reagent is measured by absorbance at 750 nm.[4] As a result, the total concentration of protein in the sample can be deduced from the concentration of Trp and Tyr residues that reduce the Folin–Ciocalteu reagent.
The method was first proposed by Lowry in 1951. The Bicinchoninic acid assay and the Hartree–Lowry assay are subsequent modifications of the original Lowry procedure.
Priest: "Universe from nothing, no laws of physics, sped up time"+ huge discrepancies. Creationism? No. Big Bang Theory
I would have assumed journals would let you err on the side of caution and simply remove your citation if it were unnecessary, but apparently citing too much can block approval.
Nowadays, most journals will expect the author to provide a camera-ready copy. They don't do any editing or typesetting anymore, they just handle peer-review and publication. Authors can modify papers following suggestions from peer review, which may include suggestions on citations. I think Nature and Science still to their own typesetting, and may commission better illustrations, but that's a rare exception. In nearly all cases where the paper has been accepted, the author has the final say about the details (within reason, of course).
Stephan
It is the field of biology that you are talking about? That's certainly not the case for my own field (linguistics). The editor still molds the submissions into a house style before it goes to the printer; the author isn't expected to do all the typesetting himself. Also, since articles are being written in English by non-native speakers, many journals will send articles on to a native English speaker to make them sound more natural (as a grad student I picked up a lot of work this way) before publication.
Probably the most commonly cited RFC is 2119, "Key words for use in RFCs to Indicate Requirement Levels", which has saved a lot of time over the years.
That RFC was only published in 1997 as well.
Citation practice (along with author ordering) is very different among the disciplines. In life sciences, there's a tendency to cite papers with lab techniques, as a shorthand for some complex procedure "The cultures were processed according to Smith[1] followed by the Jones assay[2]".
In engineering, there tends to be less of a tendency to cite a paper with methodological info: Very few people using an FFT cite the Cooley-Tukey paper; likewise, someone talking about using an ADC for sampled data isn't going to cite Nyquist, even if they say "the sampling rate was 5 time the Nyquist frequency". Likewise, in engineering, you don't see: The dice were attached to the substrate using a eutectic mixture of lead and tin as recommended by Agricola in "de re Metallica".
The editor still molds the submissions into a house style before it goes to the printer; the author isn't expected to do all the typesetting himself.
As usual those blasted physicists have it all figured out, especially in HEP. No one is expected to do typesetting: it is a process that computers excel at. Instead you provide the document source in REVTEX4 (a largeish subset of LaTeX). The journal replaces your style file with their own and it's done.
Also, since articles are being written in English by non-native speakers, many journals will send articles on to a native English speaker to make them sound more natural (as a grad student I picked up a lot of work this way) before publication.
Never seen that personally, but I'm not in linguistics. Instead, I've had (clearly non native) reviews complaining incorrectly about English constuctions that I've used, and recommending I get a native English speaker to review it. The nerve of that is quite astonishing and it would be funny if such reviewers didn't generally revel in making the life of authors as miserable as possible.
SJW n. One who posts facts.
This shows how clearly stupid citation numbers are, to compare article quality. Almost all the articles in the top 10 (and in fact, also in the top 100 from what I could tell by skimming it) are from one particular scientific sub-field: biochemistry. You'd think some of the stuff Einstein published warrant to be in the list. Or some of Heisenberg's work. Or the transistor effect. Or many other pivotal published research. But it's not there, because few have cited those articles. Well, few citations doesn't mean shit! And shouldn't be used to determine one's scientific career.
It's more common in the sciences, where the convention is to provide a LaTeX stylesheet. In subjects where manuscripts are submitted in Word format, the typesetting by the authors tends to be so bad that you need to have a professional redo it.
I am TheRaven on Soylent News
It looks like the majority of the top 20 most cited papers cover new methods or tools (e.g., a new lab technique or a new software program), not new fundamental scientific discoveries (e.g., the structure of DNA or expansion of the universe). I guess this isn't really surprising, but it is interesting. One could conclude that scientists who want to make a major impact on their field should spend their time inventing new methods for doing fundamental research and let other scientists actually do the research.
CS and math papers are more like life sciences in this regard. They cite bullshit from 30 years ago that was made completely obsolete by another paper before the author started studying the material. But [1] they [2] still [3] cite [3] anything[4] halfway[5] relevant [6] in the intro [7] paragraphs [8].
It is the field of biology that you are talking about? That's certainly not the case for my own field (linguistics). The editor still molds the submissions into a house style before it goes to the printer; the author isn't expected to do all the typesetting himself.
Ok, my experience is mostly with computer science, math and physics. Typically, you write your paper in LaTeX with a style provided by the publisher. LaTeX does the actual typesetting, of course. Some journals also have Word templates, but that's much rarer.
Stephan
If you discovered a new gene responsible for Alzheimer's you would get cited in a lot of medical journals, but devise a new and particularly useful computational method (i.e a new and particularly useful linear system solver or numerical integration scheme) and you can have an impact on nearly every scientific field.
It's a little bit different with engineering. In engineering you tend to rely a bit more on the tired and tested. FFT is a tool to get something done and doesn't need citation any more than you would need a citation when mentioning what screwdriver you used.
With sciences where you deal with less tested things the methods used aren't necessarily correct or can have flaws and limits that makes them not apply to what you are using them for. Here a citation is much more useful because it allows anyone who runs into trouble to backtrack and see what went wrong.
Its a good example. Crick and Watsons work neither discovered DNA not did it explain how the information is stored. What is actually in the paper is a model for the helix structure, which is totally irrelevant in the majority of cases that you talk about DNA.
After getting the final submission rejected 6 times. (The first failure was because it was PDF4, and they wanted PDF5 ... as if the couldn't up convert ... so I tried giving them the original source for them to use, but giving them ODF and DOC files resulted in font screwups ... so I tried generating the PDF through other mechanisms ... but they complained I had bookmarks (none of which showed up in Abobe Acrobat Professional) ... then their website said I had sent them too many PDFs (3), so I had to use their other methods ...
After spending hours on trying to get their damned website to accept my paper, I then got told by my boss that IEEE *also* makes you sign over copyright of your paper to them ... why, because you use their damned MS Word template that they can't even generate a clean paper from?
So I said fuck it, and withdrew the paper, and withdrew from the workshop (which is today) entirely. Never again will I even consider submitting a paper to IEEE.
Build it, and they will come^Hplain.
I've had (clearly non native) reviews complaining incorrectly about English constuctions that I've used, and recommending I get a native English speaker to review it. The nerve of that is quite astonishing and it would be funny if such reviewers didn't generally revel in making the life of authors as miserable as possible.
You know I have lots of friends who work with native speakers to make their texts easier to read, strange constructs is just one thing they remoev.
Typesetting isn't a process computers excel at. LaTeX is good, but not nearly as good as a good designer equipped with InDesign and loads and loads of time. It's faster and cheaper, yes, and certainly good enough for most academic journals (probably not Nature). Unfortunately, it also offers nothing (except decent typesetting) for fields that don't deal much with maths, whereas Microsoft Word offers a few nice tools, is somewhat easy to use, and has rubbish typesetting.
The article itself never mention 'data resources' that I saw, but there's a problem in many fields that the standards are to cite the 'first results' paper for that data ... for which the results portion may have already been disproved or otherwise be crap. There are a number of efforts working on being able to cite 'data' separately from 'results of the data', and in a manner that's consistent across all disciplines (as we don't know in advance who might make use of our data). You also run into problems, as the paper being cited may describe the initial release of the data, and not be useful to determine which edition was used (as that may be significant to recreate their results). See the Joint Declaration of Data Ctation Principlies, DataCite (metadata schema + DOI registry system), and the 2012 CODATA-ICSTI report, "Out of Cite, Out of Mind: The Current State of Practice, Policy, and Technology for the Citation of Data".
There are similar issues with software citation -- everyone's citing the announcement of the existing of the software, but how can you track who might've relied on a buggy version to let them know that they may need to re-run their analysis? I'm not as active in this field, but the arguments remain the same (giving proper attribution, documenting everything to make it reproducible, etc.). See the 2013 Knepley et.al paper, "Accurately Citing Software and Algorithms used in Publications" and the work of the Software Sustainability Institute (which also covers topics on writing better research software, as was alluded to in the article)
It's probably also work mentioning that our current ways of tracking 'importance' of papers are flawed. See the Altmetrics Manifesto for a collection of links to efforts to come up with other metrics and CiTO, the Citation Typing Ontology to enable a way to classify why something was cited (it might be for criticism; in most of the cases in the article, it would be "uses method in", which not all disciples feel needs to be cited).
Build it, and they will come^Hplain.
So some of the most widely used scientific discoveries never get cited.
That explains why my work did not make it to the top of this flawed metric.
Here is another link that is more readable. In a more easily readable form Candlestickmaker, S., and Helpit, Canna E. 1955, Compositio Math., 237, 476.
Giftcourt. M. F. 1956, J. Symbolic Logic, 237, 476.
Nostradamus, M. 1955, Centuries (Lyons).
Pythagoras — 520, in: Euclid — 300, Elements, Book I, Prop. 47 (Athens).
Shopwalker, M., and Salesperson, F. 1955, Heredity, 237, 476.
sed -e 's/Chuck Norris/Rajnikant/g' joke > fact
You know I have lots of friends who work with native speakers to make their texts easier to read, strange constructs is just one thing they remoev.
I'm a native speaker and I have no objection to removing strange constructs. However, the construct was not strange. It was co-authored with another native speaker and a fluent non native speaker. Neither of them could figure out what this guy thought was wrong with the construct.
SJW n. One who posts facts.
Typesetting isn't a process computers excel at. LaTeX is good, but not nearly as good as a good designer equipped with InDesign and loads and loads of time.
Not sure I agree. For something that isn't in TeX's general area then yes. However, a good designer with TeX and loads of time can do a very very good job. TeX does an excellent job of the basics like kerning (especially the modern variants which do micro kerning and stuff), spacing, breaking and so on. It also does a much, much better job than most people can do and certainly given a time budget it does a much better job than almost all people.
It's faster and cheaper, yes, and certainly good enough for most academic journals (probably not Nature).
I believe you are actually mistaken about that. I think they actually do use it internally.
Unfortunately, it also offers nothing (except decent typesetting) for fields that don't deal much with maths, whereas Microsoft Word offers a few nice tools, is somewhat easy to use, and has rubbish typesetting.
Well, it offers automatic cross referencing, with almost every conceivable variant of bibliographic styles. It also does that in a reliable and bug free manner, not something I've observed with word. Biologists spend an inordinate amount of time battling with reference managers.
SJW n. One who posts facts.
In engineering, there tends to be less of a tendency to cite a paper with methodological info: Very few people using an FFT cite the Cooley-Tukey paper; likewise, someone talking about using an ADC for sampled data isn't going to cite Nyquist, even if they say "the sampling rate was 5 time the Nyquist frequency". Likewise, in engineering, you don't see: The dice were attached to the substrate using a eutectic mixture of lead and tin as recommended by Agricola in "de re Metallica".
Yes and no. In computer science experienced authors rely on a common and rather broad knowledge base when writing their publications. There is a tradition not to cite things which are part of the common knowledge. I would not cite just because I was using the FFT, unless I was doing something out of the ordinary with it which actually requires understanding all the details of the original publication.
A very common sign of an inexperienced author is sloppy references. Typically there are too many references, and to works which are not strictly relevant for the author. When you see an introduction with 60 references outside a survey you can usually toss the paper in the trash bin and tell the authors to start over...
Always bad to reply to onesself, but [citation needed blah blah]
Here's a nice article on the problems of peer review (starting on page 2), iuncluding the problem of non native speakers trying to correct native ones:
http://www.bmva.org/_media/bmv...
SJW n. One who posts facts.
As quoted above, science is a shoulders-of-giants scheme, and I'm glad to see that the most-cited papers are not about the results of science, but rather good techniques for _doing_ science - well-considered, well-tested shoulders to stand on.
In Science, Engineering and Medicine, text book knowledge is the highest level of certainty that we have, when you graduate you're not expected to remember it all but you are expected to know how to look it up. A citation is for things that can't be found in text books, nobody seems really sure what "highly cited" indicates.
Well, few citations doesn't mean shit! And shouldn't be used to determine one's scientific career.
Having your work become (or overturn) text book knowledge in your own lifetime is by far the most impressive thing anyone in the "STEM" fields can have on their CV, it's basically a prerequisite for a Nobel prize.
And did you exchange a walk on part in the war for a lead role in a cage? - Pink Floyd.
pemerintahan daerah
to promote impact factor (and measures of citations like it) as a method of identifying the "best" science. Nature / Science / other high profile journals need to defend their positions as the most desirable publications from people who claim impact factor does more harm than good.
What about the bible? Oh, I forgot. It is not peer reviewed.
Maybe the PP works in Computer Science. In CS, it is common for authors to typeset the entire paper themselves using a style file provided by the conference or journal.
We actually *prefer* typesetting papers ourselves because our manuscripts are all in Latex anyway. Journals that want to prepare their own camera-ready copies (in particular, non-CS journals) often have trouble accepting Latex source are are accustomed to taking all of their submissions in Word...
Except for basically every journal in biology...LaTeX is very rare there, and most journals require submission in Word/RTF
That survival library is rubbish. A shitload of books about growing strawberries? What the hell? Do the people making this thing really really love strawberries that much? There are hundreds or thousands of different berries grown and eaten you all they do is collect a buch of books about strawberries. It's actually very funny, because I don't think that "library" will ever really help anyone.
I've had issues with this, although mostly with people not groking that British English is a valid form of the language. It wasn't even an American journal, it was French!
Yeah, BibTeX is more reliable than EndNote, but it's cumbersome to use and extremely poor at, well, managing references. Perhaps there are frontends with dupe control, sorting by arbitrary fields, grouping, etc., but then you're into the old Unix problem of having the choice of a gazillion applications that do one thing each, usually poorly, and with different combinations everywhere. A monstrosity like EndNote does pretty much everything you can want from a reference manager (much of it in a confusing manner, granted), which means you can get help: it's the same everywhere.
BibTeX might have most styles you can think of -- in English. When I last used it, I had to hack my own, not something I would recommend to your average researcher. Keep in mind that most of them are quite poor with computers. Asking them to use LaTeX is like asking Unix programmers to think of the end user: it's neither relevant to them, nor helpful.
Thank you, Melinda French Gates . . .
the authors failed to cite critical research (mine), reject for publication -- reviewer 3
Yeah, BibTeX is more reliable than EndNote, but it's cumbersome to use and extremely poor at, well, managing references. Perhaps there are frontends with dupe control, sorting by arbitrary fields, grouping, etc., but then you're into the old Unix problem of having the choice of a gazillion applications that do one thing each, usually poorly, and with different combinations everywhere.
How do you mean? Is this for some sort of display purpose other than in the bibliography of the paper? BibTeX is mostly just the database and tools for turning that plus a document into a bibliography. Beyond that it doesn't do any "management". You just have a text file with every paper you've ever referenced in it.
Though I spent over 10 years in academia during which time, my bib file got longer and longer. I don't think I ever had a problem managing it, nor with duplicates.
BibTeX might have most styles you can think of -- in English.
Yeah, I've never used anything other than English. In terms of styles I've never encountered a journal for which I couldn't find an appropriate style, especially as natbib allows quite a lot of flexibility from within a document.
SJW n. One who posts facts.
Nowadays, most journals will expect the author to provide a camera-ready copy. They don't do any editing or typesetting anymore, they just handle peer-review and publication.
I've never had a journal require a camera-ready copy. In fact, I think that concept may be archaic enough that Mr Shultz means something completely different. Honestly, I don't even think modern journal publishers use a camera anymore.
Every manuscript I've submitted in the past 15 years (biological sciences) has been an MS Word .doc file (some journals will let you do LaTeX, but most are confused by such things). Those journals may not spell-check, or grammar-check, or change any of the words in any way, but they very definitely handle the layout. Every manuscript I've submitted or read has been set in 12-point font, double spaced, single column. Usually with figures as separate pages at the end. Most articles I read are 9- or 10-point font, single spaced, two column, with text wrapped around figures.
Nothing of his seems to have made the list. Sort of surprising.
CUR ALLOC 20195.....5804M
Out of curiosity, what journals are you submitting to that require "camera ready" copy? I'm aware of very few in the life or physical sciences, and most of those aren't exactly top tier.
Most journals expect the text (including citations) in a "standard" format, I'm aware of none that won't accept any semi-recent version of word (.doc/.docx), most accept PDF, many will accept RTF, a few will accept TEX (maybe most if your field is physics or math). They generally want each figure as a separate file, either vector or bitmap with a fairly high minimum resolution, so they can resize the images and reflow text around them. How tables are presented/accepted is pretty journal specific, but this is the one area where many journals may reformat your work.
As for Nature and Science, I "created" the cover image for a supplemental issue of Nature Structural Biology quite a few years ago. For covers and promotional things, their art department gets involved and the final image may only look vaguely like what they were sent in the first place. Really, I probably could have sketched something on a napkin instead of spending time trying to make a decent figure in the first place with the changes they made in the end (though we did get to give our approval for the final image). As for the associated review article, the figures were all published as submitted; one of the editors may have asked for a change, but we would have made it ourselves.
Well you know how the French and British get along.
http://dl.acm.org/citation.cfm...
New Economic Perspectives
Pythagoras — 520, in: Euclid — 300, Elements, Book I, Prop. 47 (Athens).
The citation is probably wrong. Although we don't know exactly where Euclid's Elements were written, he lived in Alexandria and not in Athens. The oldest known complete edition was also edited in Alexandria.
Yes, I'm a citation nazi.
As in any field, being different is not always good
The standard practice in molecular biology/biochemistry/genetics is to cite only recent papers, unless there is a real need for an older one
The Watson Crick paper is so old, that it would be odd for someone to cite it; this would probably raise a flag in the reviewers mind - whats going on here
And, of course, you never want to raise any doubts in the minds of the reviewers
ymmv
and read a lot of others in the field of molecular biology/genetics/cell biology/biochem
A lot of times, the citations are just awful: The author will cite ref 1 as a method, and ref1 will say according to the procedure of jones, and jones will say method to be published (I've actually had this happen)
The bio lit is so disorganized, and so large, that almost everyone misses stuff
etc
I mean, who cares about the citations ? unless you do something really dumb, all people care about is the quality of your science
It took me a long time to realize that science, at least in mol bio, is *not* scholarly; it is a who is 1st competition for the most newest thing..scholarly stuff like citations are just not that high a priority for a lot of people
I've published a few times as a Biologist. At least for our Biology journals the editor does to a small amount of editing. I do not know of a journal that make you do your own typesetting.
How do you mean? Is this for some sort of display purpose other than in the bibliography of the paper? BibTeX is mostly just the database and tools for turning that plus a document into a bibliography. Beyond that it doesn't do any "management".
Exactly. The ability to view the database, sorted in any order imaginable, or ordered into groups, either manually or through live searches. It's a very useful tool when writing a review with hundreds or references, and is nice to have even if you've got just a few dozens.
Most physics journals I've submitted to, even meeting proceedings that were not putting much effort into their proceedings, will convert the latex files into some other type setting program. They explicitly warn of this in some cases, and in other cases it is kind of obvious, where after the peer review process, they won't accept latex submissions for the layout draft. At that point you need to tell them what to change, and they change it in whatever other program they use.
Very few people using an FFT cite the Cooley-Tukey paper;
And yet this is by far one of the most cited math/CS papers...
Exactly. The ability to view the database, sorted in any order imaginable, or ordered into groups, either manually or through live searches. It's a very useful tool when writing a review with hundreds or references, and is nice to have even if you've got just a few dozens.
Hm, maybe I live is a web search world, but I never found myself wishing for that kind of thing. It tended to be I'd read a paper finding it through web searches, on the website of a researcher I knew to be be important in the field, or cited elsewhere.
If I wanted to cite it, I searched the bib file for the author name or the title (that was always trivial enough) and if present, I added it to the document, otherwise I made a new entry.
I never found myself browsing through the database, or wishing I could order by various things so I could browse through lists.
Aside from a thesis, which obviously has a long list, I have had papers go into the hundreds.
SJW n. One who posts facts.
Hm, maybe I live is a web search world, but I never found myself wishing for that kind of thing. It tended to be I'd read a paper finding it through web searches, on the website of a researcher I knew to be be important in the field, or cited elsewhere.
But that's just not practical when you want a current overview of a huge field. With EndNote, you typically dump the entire list of search results from the database, and then start reading abstracts (included in the reference file), sorting relevant from irrelevant, and then download PDFs to read (which are then stored along with the references). It's a research tool and a retrieval tool. BibTeX isn't.
But that's just not practical when you want a current overview of a huge field.
How huge us huge? It was fine for writing a thesis. It was fine for writing some decently long literature reviews.
With EndNote, you typically dump the entire list of search results from the database, and then start reading abstracts (included in the reference file), sorting relevant from irrelevant, and then download PDFs to read (which are then stored along with the references). It's a research tool and a retrieval tool. BibTeX isn't.
Not to put too fine a point on it, but that's what I used google for. Search the web, find papers, read abstracts then papers. If I decide to cite them, I stick the reference in the bibtex database.
SJW n. One who posts facts.
You use Google for research? Seriously?
You use Google for research? Seriously?
How is that worse than using endnote?
And yes, I do. They have this quite useful tool called google scholar. For many papers you can easily follow chains of references, reading the papers citing the ones of interest, and easily looking up the ones cited. They also often have links to open PDFs so you don't have to bugger around trying to find how to get access to some journal or other which won't authenticate properly.
But yes, these days google features prominantly in any literature searches I do. I can easily find the websites of the relevant academics. I can find papers, citing papers, cited papers and all sorts of useful things. Seriously, who doesn't use google for that sort of thing?
SJW n. One who posts facts.