Wikipedia talk:WikiProject Chemicals#user:Plasmic Physics

File:Information.svg

An editor has requested for :Alpha hydroxy acid to be moved to another page. Since you had some involvement with Alpha hydroxy acid, you might want to participate in the move discussion (if you have not already done so).

In [https://en.wikipedia.org/w/index.php?title=Flavin_adenine_dinucleotide&curid=1491100&diff=1264708353&oldid=1247491485 edit] this edit], User:Graeme Bartlett says that 3dmet is dead. Indeed, links such as [http://www.3dmet.dna.affrc.go.jp/cgi/show_data.php?acc=B04792] are timing out for me also. Anyone know if this is a transient problem or else if we should remove this infobox item across the wiki? DMacks (talk) 04:20, 23 December 2024 (UTC)

:3DMet has been dead for years now. So I have not been able to confirm any values to add. Since it is dead, our 3dmet links are no longer useful. So I was thinking that I would remove the parameters when I come across them in the chembox. My current effort is to add missing ChemSpider entries. But as I go I might add other values or change wikidata link if wrong, (or remove 3DMet). Back in 2018 I contacted Miki Maeda from National Agriculture and Food Research Organization about 3DMet, But I have not done that since the recent multiyear outage. Graeme Bartlett (talk) 05:51, 23 December 2024 (UTC)

::Sounds like we should turn off its display ASAP (no sense giving readers a broken link). And also set up an editorial note and hidden tracking category for its use (445 pages, by a quick heuristic). Eventually a bot could simply remove them, but for now at least we'll be aware of it when we edit or if someone gets bored and wants to gnome it. DMacks (talk) 05:56, 23 December 2024 (UTC)

:::I did a quick and easily-reversible turn-off of the display,[https://en.wikipedia.org/w/index.php?title=Template:Chembox_3DMet&diff=prev&oldid=1264773621] pending stronger consensus for removing it as a supported template field altogether. DMacks (talk) 12:52, 23 December 2024 (UTC)

::::Time to kill {{tl|Chembox 3DMet}} and remove its fields from the articles? For anyone who deals with wikidata, there is the {{Wikidata property link|P2796}} property. Should it be annotated as being dead somehow? DMacks (talk) 01:50, 20 March 2025 (UTC)

:::::I contacted the developer, Miki Maeda about it. Apparently the Internet available database disappeared in a server migration. It may come back on line within 3 years as it was supposed to be publicly available. Perhaps there are other options such as incorporating the info into Wikidata, but for the next year or so don't expect to see any publicly accessible system for 3DMet. The data is available for individual researcher use on request. Graeme Bartlett (talk) 11:46, 25 March 2025 (UTC)

Anyone willing to provide a standard structural formula? I'm currently not sitting at the right PC to do it myself. --Leyo 19:46, 31 January 2025 (UTC)

:I replaced it with a black and white SVG image. Innerstream (talk) 20:24, 31 January 2025 (UTC)

:: Thank you. --Leyo 22:11, 2 March 2025 (UTC)

I just cited a few-paragraph ditty on glyceraldehyde from American Chem Soc's archive of "Molecule of the Week" (MOTW). The archive https://www.acs.org/molecule-of-the-week/archive.html?archive=All has hundreds of entries. Some entries are so short that they are not very useful, like abacavir at https://www.acs.org/molecule-of-the-week/archive/a/abacavir.html, Many, like the one for glyceraldehyde, are multi-paragraph commentaries. The MOTW site is open access, and the discussion is mid-level such that these articles could enhance Wikipedia articles. Like all things with ACS (a for-profit, unlike RSC), these links are a come-on for using SciFinder. --Smokefoot (talk) 16:37, 13 February 2025 (UTC)

Posting here instead of the WP:CHEMVAL talk page to get more comments. The situation right now is that almost all chemical articles on Wikipedia link to a page with outdated information about chembox validation. On that page it still says there's a bot that updates validation based on an index (which was not linked on the page until I added it yesterday). As of now, CheMoBot has been inactive for 7 years, so no one is keeping an eye on the File:Yes check.svgYs and File:X mark.svgNs. I marked the page as historical which an admin quickly reverted so I assume that was the wrong thing to do. How should we go about updating the page so people know chembox validation doesn't work like it used to? HansVonStuttgart (talk) 09:20, 23 February 2025 (UTC)

:There was a discussion of chembox validation back in 2020, now at WT:WikiProject Chemistry/CAS validation#Chembox Validation, CheMoBot. I think that such bot-handled validation is not useful and we should scrap it, along with the ticks and crosses. We can rely on human validation, as for the vast majority of edits. The bot was validating what are provided by clickable links to external databases, so human verification is usually easy. Mike Turnbull (talk) 14:55, 24 February 2025 (UTC)

::I've always found them to be useless. Wasn't there a plan to have wikidata handle this? Project Osprey (talk) 15:00, 24 February 2025 (UTC)

:::Yes, that was part of the 2020 discussion. The CAS Common Chemistry db became available then. Wikidata is a much better place to do any validation. Mike Turnbull (talk) 16:11, 24 February 2025 (UTC)

::::So what we need them is a report that shows differences to Wikidata. But then once checking the entry we should be able to fix or explain the difference and record it. (eg family vs specific) ( or stereoisomer) (error in database entry) (multiple database entries for one thing). Then it does not have to be rechecked over and over.

::::I have been slowly manually checking CAS numbers that did not have a tick or cross and adding the green tick, but is that a waste of time? Most of our readers would not care. Some of our readers may be happy to know that the entry was checked, but for the very serious ones, they should check it for themselves. Perhaps we need some categories that are useful to us to identify discrepancies. Graeme Bartlett (talk) 00:13, 25 February 2025 (UTC)

:I removed the references to CheMoBot on the page for now. It's still not good enough to be a real explanation of the process: it now implies people are doing exactly the same tasks CheMoBot used to do, but that's closer to the truth than claiming the bot is still active. If it ever comes back online, my changes can be reverted. HansVonStuttgart (talk) 08:36, 25 February 2025 (UTC)

If you go to [https://echa.europa.eu/universe-of-registered-substances this page] and scroll to the bottom you can download a list of all chemicals registered in the EU (you must check the disclaimer box). 4463 of these have active registrations for >100 tons per year. Is there a way to cross-reference these (by CAS number?) against our pages to generate a list of things which are

:# Produced on a significant scale:

:# We don't have pages on.

It would strike me as very good worklist to have. Project Osprey (talk) 10:10, 28 February 2025 (UTC)

:Here's a dump of the wikilinked chemical names: User:Marbletan/REACH. Not sure it helps, but at the very least it might be a place to get ideas for new articles (or new redirects to existing articles). Marbletan (talk) 15:00, 28 February 2025 (UTC)

::Thanks for that. Browsing through, there are significantly more red links than blue. Some of that might be false negatives - ECHA tends to use formal IUPAC names. --Project Osprey (talk) 15:37, 28 February 2025 (UTC)

:::There are also loads of redlinks that we would never want to have articles on, e.g. "Reaction product of....", "No public name....", "Slag....". The list will be very useful if pruned of the ones of obvious no interest. Note that any name having square brackets breaks the wikilinking. Thanks, User:Marbletan for producing that so rapidly. Mike Turnbull (talk) 15:59, 28 February 2025 (UTC)

::::... incidentally, your list is all 22,000+ on the REACH list. I think we only need to focus on the >100 ton examples, per Project Osprey's original suggestion. Mike Turnbull (talk) 16:06, 28 February 2025 (UTC)

:::::Thanks for catching that. I have now trimmed the list to those that are denoted as ">100 tpa". Much shorter and manageable. Marbletan (talk) 16:18, 28 February 2025 (UTC)

:::::Please feel free to edit the page as you see fit, to prune it of "ones of obvious no interest" or to fix the broken wikilinks. Marbletan (talk) 16:19, 28 February 2025 (UTC)

::::A CAS look-up would probably be more helpful, but I've never been able to get wikidata to work for me. The reaction mixtures can be deceptive "Reaction products of phosphoryl trichloride and 2-methyloxirane" is Tris(chloropropyl) phosphate. This reflects ECHA role as a regulator, they wont use the simple name because the material is never produced pure. Sometimes that can get interesting, there are a few compounds that they refuse to use the CAS numbers for because the CAS relates to a pure compound. Frankly, even if only 1000 of these warrant pages that's still enough to keep me going for years... Project Osprey (talk) 16:53, 28 February 2025 (UTC)

:::::I've deleted about 1,000 that would be highly unlikely to be names of articles in Wikipedia and begun to put the list in alphabetical order. Your point about reaction products is reasonable but we should only ever have them as the actual product name. Mike Turnbull (talk) 17:27, 28 February 2025 (UTC)

::::::Oh yes, definitely. I wouldn't even use such descriptions as redirects, I'm just trying to explain why they exist at all. Project Osprey (talk) 17:42, 28 February 2025 (UTC)

:I have started making some redirects. But some do not appear worthwhile eg Aluminum, (octadecanoato-.kappa.O)oxo- which we have as Aluminium monostearate. So Marbletan, would you like me to note these on your list? Or do it here? Graeme Bartlett (talk) 23:00, 28 February 2025 (UTC)

::In the cases where a page exists, either on Wikipedia or Wikidata, but the redirect doesn't seem appropriate, it seems like it would be a good idea to note it on the list in some way (but actual format probably doesn't matter much). Marbletan (talk) 17:10, 2 March 2025 (UTC)

:It would also be helpful to cross-reference them if corresponding Wikidata pages exist.--Leiem (talk) 03:02, 1 March 2025 (UTC)

::There are quite a few Wikidata entries for pure substances when there is no article here. Graeme Bartlett (talk) 03:39, 1 March 2025 (UTC)

:::I'd expect most of the pure substances to have Wikidata entries. Note that you can link to them using the template {{t|ill}}, which will give redlink entries like {{ill|1,2-dimethylimidazole|qid=Q27286970}} (I've updated that one in Marbletan's list as an example). This makes is much easier to create chemboxes when writing articles, as, of course, Wikidata contains many of the IDs we usually include, as well as InChI, SMILES etc. When such links occur in mainspace lists (e.g. List of herbicides), there is even a bot that comes along to convert these links into standard links when an article is created, as you can see in the edit history for herbicides. Mike Turnbull (talk) 12:33, 1 March 2025 (UTC)

I will paste some code I have used to process the data (no guarantee of working though):

Convert the csv ECHA file to tsv, remove verbose content to make list of name, EC, CAS, (amount indicator 3 is >100 tpa), infocard

: sed -e 's/\(.*\),[^,]*,[^,]*\/\([0-9.]*\),[^,]*,[^,]*$/\1\t\2/g; s/,intermediate\t/\t2\t/g; s/,<100 tpa\t/\t1\t/g; s/,>100 tpa\t/\t3\t/g; s/,not yet assigned\t/\t0\t/g; s/,active registrations(s) under REACH[,]*\t/\t/g; s/,\([0-9\-]*\),\([0-9\-]*\)\t/\t\1\t\2\t/g; s/^"//g; s/"\t/\t/g' output

retrieve Wikidata entries with infocards returning QID, label, CASNo. infocard, and EC_number:

:curl --data-urlencode "query@wiktquery" -H "Accept: text/tab-separated-values" https://query.wikidata.org/bigdata/namespace/wdq/sparql >outputwikidata

THe file wiktquery has this content:

SELECT ?item ?itemLabel ?casNo ?echaId ?einecs ?article WHERE {

?item wdt:P2566 ?echaId .

OPTIONAL { ?item wdt:P231 ?casNo . } # CAS Registry Number

OPTIONAL { ?item wdt:P232 ?einecs . } #EINECS

OPTIONAL { ?article schema:about ?item .

?article schema:inLanguage "en" .

?article schema:isPartOf .

}

SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }

}

Graeme Bartlett (talk) 01:14, 5 March 2025 (UTC)

Just wanted to point out that finally, all the Chemicals articles in the wikiproject have been assessed. You can view the classification in the big chart on the main page. This is essentially thanks to a combination of the automatic classification of set index articles as List-class added to Module:WikiProject Banner Shell, and people manually classifying articles. Just got the last article classified. Mrfoogles (talk) 21:33, 7 March 2025 (UTC)

{{ping|Lamro}} Lamro, do you really think that "Technical Paper. U.S. Government Printing Office. 1917" and "American Druggist and Pharmaceutical Record. ...1895" are good foundations for something (Ammonium oleate) notable? --Smokefoot (talk) 23:16, 7 March 2025 (UTC)

:Actually those sources have substantial content, and would be reliable. But they are not suitable for how they are used, as there is no chemical formula, and they are both on applications of the substance. Graeme Bartlett (talk) 00:26, 8 March 2025 (UTC)

When a new Wikipedia article is created about a chemical, is there any bot or other tool that looks for an existing Wikidata entry for that substance, creates it if not existing, and links the wp article to it? It's not a hard manual task to "look for, and link if existing" (and then flag if need to create). And I'm glad to see a bunch of editors creating new chemistry/drug articles. But that is leaving a lot of missing wikidata links. Would it be useful as a quick step to have chembox (and drugbox too) populate a category for missing wikidata links? DMacks (talk) 14:44, 12 March 2025 (UTC)

:wikidata:User:Pi bot does this, it just takes a while. I've seen it happen on a couple of articles I've created, like Kuicaoxi. Reconrabbit 16:17, 12 March 2025 (UTC)

:When new articles show on the list, and the PubChem entry links to Wikidata then I will link the page to Wikidata. Quite often though there are duplicate Wikidata entries, so I then try to merge. Graeme Bartlett (talk) 23:44, 12 March 2025 (UTC)

Yesterday, a new editor, User:Bram Lentjes and an IP made a very large expansion of the article on VG. These edits raise several issues which I have not followed up in detail (e.g. [https://en.wikipedia.org/w/index.php?title=VG_%28nerve_agent%29&diff=1280620845&oldid=1280613206 use of predatory journal as source]). What I want to discuss first is [https://en.wikipedia.org/w/index.php?title=VG_(nerve_agent)&diff=next&oldid=1280630841 this edit], which introduced a section saying {{tq|We acknowledge the use of AI assistance, specifically ChatGPT, to help paraphrase and refine portions of this text. While the final content was reviewed and edited by our team, AI was utilized to enhance clarity and readability.}} I have no idea who "our team" refers to and since this is an IP we can't even properly interact with them (although I will leave a message on their talk page). I don't believe there is a consensus that it is acceptable to use large language models in articles, far less to add such acknowledgements. Comments? Is there another venue where this needs to be discussed? Mike Turnbull (talk) 14:06, 16 March 2025 (UTC)

:... just one very obvious flaw in [https://en.wikipedia.org/w/index.php?title=VG_(nerve_agent)&oldid=1280788122 the current version of the article] is the discrepancy between the rat oral toxicity figure of 3300 mg/kg in the "Toxicology data" section and the 5.4 mg/kg value in the chembox. My inclination is to revert back to the [https://en.wikipedia.org/w/index.php?title=VG_(nerve_agent)&oldid=1273946963 version of 4 Feb 2025] and try again. Mike Turnbull (talk) 14:44, 16 March 2025 (UTC)

::I think there's something similar going on in Draft:Norlichexanthone, which is the result of a university group project (see discussion on my talk page for more). Esculenta (talk) 15:39, 16 March 2025 (UTC)

:::Support. Revert VG (nerve agent) and ask for stepwise edits, followed by appraisal by experienced editors. Thank you for identifying this issue. --Smokefoot (talk) 15:48, 16 March 2025 (UTC)

:::I declined Draft:Norlichexanthone. DMacks (talk) 02:21, 19 March 2025 (UTC)

:Support. Also suggest reversion to February 4th. Bizarrely searching google for one of the IP addresses used (213.124.171.69) now brings up the VG page as the top result. Anonrfjwhuikdzz (talk) 16:19, 16 March 2025 (UTC)

::After further investigation, I have reverted all these edits. I will now work to improve the article but that will take time as it needs a complete rewrite. Mike Turnbull (talk) 12:09, 18 March 2025 (UTC)

:I just identified an upload from User:Bram Lentjes on commons as attribution and license failure (claimed own, but is dup of another CC file). So either they are using AI for more than just grammar and tweaks of their own content (and in the process cut'n'pasting AI's failure to cite its own directly copied content) or their original writing itself has a shadow of potential plagiarism over it. DMacks (talk) 04:41, 19 March 2025 (UTC)

::Hi DMacks, could you elaborate on which picture this is? I hope I can shine some light on what went wrong and help fix it. Newtonpersquaremeter (talk) 15:00, 23 March 2025 (UTC)

:Hi Mike, I saw your comment and decided to make an account to contact you. As user Esculenta suggested, the edit on the amiton page was also the result of a group project for our course toxicology. I'm sorry to hear that our edits are not according to Wikepedia's standards and I would like to help improving our work to make it publishable again. About the aknowledgement: Including a statement about the use of AI in our project would result in bonus points for the project. And as the aknowledgement states, I (and as far as I know the rest of my group) only used AI to paraphrase. All the research done and sources found were done manually, without the use of AI. Could you please elaborate on the use of predatory journal as source (and other problems)? Newtonpersquaremeter (talk) 14:59, 23 March 2025 (UTC)

::Here are some comments from a regular editor here: most or at least a lot of homework assignments are substandard. Homeworkers are often undergrads who barely comprehend chemistry. Their work is generally unsupervised or lightly supervised by teachers who have no experience contributing content to Wikipedia. Wikipedia is not a platform for students/teachers who want to feel good about classroom projects. Good sources? Reviews in highly ranked scholarly journals, chapters from iconic textbooks. See WP:SECONDARY, WP:TERTIARY.--Smokefoot (talk) 15:18, 23 March 2025 (UTC)

:::To be fair, I am unqualified in chemistry and have written some fairly decent articles (I don't know nothing, but that's about it). A lot of it is just understanding how to write for Wikipedia, which the education program helps with Mrfoogles (talk) 04:05, 27 March 2025 (UTC)

::@Newtonpersquaremeter there was so much wrong with the version of the article which I linked above that I hesitate to point out all the problems. To take only the most obvious: 1) The toxicology section. Pubchem is not really a source: it is a database that collects together information published elsewhere. The editor who added the figure of 3,300 mg/kg for the LD₅₀ had in any case misread pubchem, which has the value as 3,300 ug/kg (i.e. a factor of 1,000 more toxic) and should have quoted the original source which pubchem provides. Extrapolating this to humans is fraught with problems and should never be done here in Wikipedia without a cast-iron source: none was provided. I used the secondary source from the ICI book, which states the rat LD₅₀ as "about 5 mg/kg", which is close enough: the point is that this value is similar to that for parathion, a widely use material at that time. Amiton was, as far as I am aware, never marketed and over-emphasis on its toxicity is irrelevant to its notability as the first V-agent.

::2) A more egregious error was to use doi:10.13188%2F2328-1723.1000019 as a "source" for the Adverse Effects and Toxicity section. That is a predatory journal and, more to the point, neither it or the other cited source for that section even mentions amiton: which is not surprising since as I've stated, it was never marketed!

::3) Now to the use of doi:10.1080/10426507.2018.1540491 as a "source" for the chemical synthesis section. You editor can't have read that paper, since it doesn't mention how amiton is made: it is a brief account of how it can be assayed for acetylcholinesterase activity. I suspect that this editor made up the information from general knowledge of organic chemistry and then added that citation as a plausible place which might have had the information. Sadly, it doesn't. There are proper ICI sources for the synthesis, which I have now included.

::It is completely unacceptable to me that any set of editors should use Wikipedia as an outlet for coursework without rigorous oversight from experienced supervisors in order to guard against the sorts of errors I have mentioned. Anyone who thinks that LLM are appropriate at their present state of development for writing Wikipedia articles is, in my opinion, seriously misinformed. Mike Turnbull (talk) 15:55, 23 March 2025 (UTC)

:::@Michael D. Turnbull I took some time to dig into these problems.

:::1) the problem with the 3,300 is caused by a mistake of a group member. In the Netherlands (where we live), the , is used as a decimal separator instead of the . Therefore this mistake was made out of habit, and the wrong notation was used. I do agree with you that they should've used the source cited by pubchem. About the extrapolation to humans: I have not done this research, so I cannot comment on what they exactly did there.

:::2) We (or at least I) didn't know this was a predatory journal. I didn't write this section either so again i cant comment on why this happened.

:::3) This is a mistake on my side; I did the reference list, and to make the refences I used the automatic citation tool. The tool grabbed a different paper from the same author, and I did not check if it was correct, the authors matched. As seen in the edit [https://en.wikipedia.org/w/index.php?title=VG_(nerve_agent)&oldid=1280610602 14:34, 15 March 2025], the actual source is Althoff MA (2018). Karaghiosoff KL; Klapötke TM (eds.). Preparation and analytical investigation of Amiton and Amiton-like compounds closely related to the Chemical Weapons Convention, instead of Althoff, M. A.; Unger, C. C.; Bützer, P.; Metzulat, M.; Klapötke, T. M.; Karaghiosoff, K. L. (2019-02-07). "Bioactivity and toxicological study of Amiton and related isomers". This source does contain the information about the synthesis of Amiton. This fault is on me and next time I will make sure all sources are checked.

:::I do agree with you that for this project to be benificial for Wikipedia, more supervision is needed. Also, I agree that LLM's are not appropriate to write Wikipedia pages. I believe this isn't the professors stance either; I think the idea behind the bonus point for AI aknowledgement is that they know they cant make sure we don't use it, so it is better to actually properly reference to the AI used. Newtonpersquaremeter (talk) 20:56, 23 March 2025 (UTC)

::::@Newtonpersquaremeter: a couple of points based on your reply. The source you mentioned above for the synthesis is [https://edoc.ub.uni-muenchen.de/22609/1/Althoff_Marc_Andre.pdf this thesis] by Althoff. We tend to avoid theses as direct sources since they are not peer-reviewed and can be error-prone. They can be good sources if one takes the trouble to drill down into them and seek out the underlying literature. Althoff does cite doi:10.1039/JR9600000637, which is what I have now used for the synthesis section. We have guidance on using LLM: see the essay WP:LLM and its talk page. There is no consensus for their use but they are not forbidden by policy. I personally would never use them but the main point is that anyone who does is supposed to say so in their edit summary, not in the actual article. Thanks for being prepared to discuss these issues: Wikipedia can only be successful if editors communicate. Mike Turnbull (talk) 11:42, 24 March 2025 (UTC)

:::::@Michael D. Turnbull Thank you for your reply. I'll keep these things in mind for the future. And no problem, I'm glad I could help! Newtonpersquaremeter (talk) 19:03, 24 March 2025 (UTC)

File:Information.svg There is a requested move discussion at Talk:Ergoamides#Requested move 23 March 2025 that may be of interest to members of this WikiProject. TarnishedPath^talk 03:22, 7 April 2025 (UTC)

I have nominated for deletion an article, arabilin, that may be of interest to members of this Wikiproject. See Wikipedia:Articles for deletion/Arabilin. Innerstream (talk) 12:58, 16 April 2025 (UTC)

At Tellurophenes, an article with an unusually large amount of detailed content, there is a requested move Tellurophenes → Tellurophene which could use the input of knowledgeable editors. Please comment at Talk:Tellurophenes#Requested move 4 May 2025 if interested. Adumbrativus (talk) 04:56, 14 May 2025 (UTC)

Fluspidine popped up as a new article, listed as PubChem [https://pubchem.ncbi.nlm.nih.gov/compound/71719166 CID 71719166]. PubChem calls that entry "[18F]-Fluspidine", which matches what our article's chemical is (both content and the cited ref). But that database's other mechanically derived entries (IUPAC name, SMILES, etc.) are instead specifically the ¹⁷F isotopolog. ChemBL [https://www.ebi.ac.uk/chembl/explore/compound/CHEMBL2314421 CHEMBL2314421] likewise has contradictory isotopic details within this entry. Should we include these database tokens in the infobox and flag them somehow, or omit them and include a note explaining why? DMacks (talk) 14:09, 14 May 2025 (UTC)

:I'd leave the IDs out until the databases catch up, as they are likely to do. Meanwhile User:Boghog might like to update the .svg image file to show the correct isotope! We can, of course, alter the IUPAC name and SMILES we use as these are "sky-is-blue"-type information that doesn't need a specific citation. Mike Turnbull (talk) 17:01, 14 May 2025 (UTC)

:: :File:Fluspidine.svg {{fixed}}. Good catch DMacks. Boghog (talk) 18:23, 14 May 2025 (UTC)

:::That was me! I've reported the error to Chemspider for their #29397143, so hopefully that will get fixed. Mike Turnbull (talk) 10:18, 15 May 2025 (UTC)

:If there is a correct entry eg [https://www.ebi.ac.uk/chembl/explore/compound/CHEMBL1645202 CHEMBL1645202] then please use that instead of a wrong or narrow one. But in this case I would put in a wikitext comment to state the wrong pubchem entry and what the problem is. That should stop someone else putting in the incorrect entry. Another common type of error in PubChem is a charge error, particularly for transuranic compounds. Graeme Bartlett (talk) 23:11, 14 May 2025 (UTC)

::The CHEMBL1645202 link goes to a page for the parent compound, presumably the 19F (i.e. normal) version. That page has a box lower right which links to CHEMBL2314421, describing it as the 18F compound but with the wrong drawing! Mike Turnbull (talk) 10:26, 15 May 2025 (UTC)