Remember When?

Remember how we used to generate research reports?

"The scientist in a pharmaceutical company prepares his compound documents, using a variety of text and graphics systems, in a laborious cut-and-paste procedure, and transmits a hard copy to a records center, where photocopies and microforms are made and indexing takes place. When the end-user wants to retrieve his report, he asks an intermediary to do a search. The result will be a report number or access address. The hard-copy report then has to be manually retrieved, perhaps copied or printed, and delivered by mailman to the end-user. . . it is also evident that research end-users are not accessing full-text online databases to any large extent."

 - From J. Chem. Inf. Comput. Sci., 1991

It's truly amazing how quickly technology has accelerated online publication in the intervening two decades.

Public library lintel, USA

Update (Aug 3) - Derek Lowe weighs in:

@SeeArrOh Yep, that's how my continuation exam in grad school was done. Typewriter, plastic template, rub-on letters. Just before @Chemdraw.
— Derek Lowe (@Dereklowe) August 4, 2014

Fukuyama Mea Culpa in Organic Letters Corrections

Remember all the recent dust-ups over potentially doctored NMR spectra? How about when Organic Letters EIC Amos Smith went on the offensive, hiring a full-time data analyst to sniff out fishy details?

Well, we have a major opening salvo in the war on sketchy Supporting Information (thanks to an anonymous JLC commenter for the tip-off). Seems that Prof. Tohru Fukuyama (U. Tokyo) has issued six simultaneous Org. Lett. corrections in yesterday's ASAP section. (1, 2, 3, 4, 5, 6).

This is a big deal, not least because Prof. Fukuyama's reputation looms large in the synthetic community, but more because the corrections impact challenging targets from his group (manzamine A, huperzine A, lysergic acid, histrionicotoxin) and span multiple years (2008-2013).

Shockingly, the corrections all admit deletion of solvent peaks or impurities from final, published spectra. This correction announces 20 such "edits." In some instances, when the group re-examines the spectra, they're so unreliable that they cannot correct the record: "Samples included some unknown impurities, thus the exact purity could not be determined."

I'm not sure whether to applaud or condemn the Fukuyama group - if they chose to come forward voluntarily to correct the record, I applaud the effort. However, if they were instead strong-armed by the OL Data Analyst, then a culture of data manipulation has been uncovered at the highest levels of our field (reminder: this is just one journal, hold on for the slate of corrections in other venues).

Wish I could say that this meant "Case Closed," but I think we're only seeing the tip of the corrections iceberg in Organic Letters.

Friday Fun: How to Fund Your Data Analyst

Remember Amos Smith's Editorial, discussed here yesterday?

(and here, and here, and here)

I wondered, on Twitter: How many submissions does Organic Letters get in a year, anyway?

Sonja Krane, a JACS editor, set me straight:

@seearroh Journal submission statistics aren't publicly available. Sorry I can't be more helpful in this case!
— Sonja Krane (@sonjakrane) June 6, 2013

Rats, foiled again! But then, an interesting tidbit from Stu Cantrill over at Nature Chemistry
(N.B. Stu used to work at OL):

@seearroh When I was there, accept rate was getting down to 50%. I suspect now 30-40%, so look at papers published in 2012 & go from there.
— Stuart Cantrill (@stuartcantrill) June 7, 2013 

Hmm, so all I have to do is count. In 2012, Organic Letters published 24 issues, which seem to have an average article count ~80 / per.* So that's 2,000 articles / year, give or take 100. Now, let's assume Stu's lower range (30% acceptance) - that's 7,000 submissions. Back of the envelope, I'd guess an average Supporting Info section to clock in at around 40 pages nowadays.

That's 280,000 pages of SI.
Pity the poor Data Analyst.

But...what a great way to FUND this potentially burgeoning "alternative" career! A nominal fee of, say, $0.10 / SI page - price of a photocopy from way back, kids - would immediately bring $30K into the journal's coffers. A $3 "data verification" fee per manuscript brings another $21K. Not big money, but we're now into the realm of serious subsidy for someone's salary.


Readers: Would you pay $7.00 to submit your OL manuscript?

* [(Dec 21 + July 6 + Jan 6 + Apr 20) - (corrections + editorials)] = 318 articles / 4 = 79.5