Template talk:Infobox drug/Archive 4

New chemical formula style
I'm going to do the same thing (create a category) for pages that use the old chemical formula style. (Category:Drug pages needing formula fontification. -Selket Talk 19:42, 24 February 2007 (UTC)
 * That is excellent. I can't believe there are only 45, though. Fvasconcellos 19:47, 24 February 2007 (UTC)
 * Because the category is transcluded, it only gets added when the page is rendered. We need to wait until people view all of the pages in question.  More are popping up. --Selket Talk 19:58, 24 February 2007 (UTC)
 * OK, thanks. I don't know whether you've noticed, but some extra whitespace is now being added to the top of articles using the new formula style. Might it have anything to do with the new category? Fvasconcellos 23:32, 24 February 2007 (UTC)
 * Yes it might. I'll take care of it.  Also, I've begun fixing them with AWB, but I don't know what to do with the formulas that have congugated acids (e.g., Amiodarone: C25H29I2NO3 . HCl).  Are we moving to a purely empirical formula for these, or should I leave them?  What's the consensus?  -Selket Talk 23:47, 24 February 2007 (UTC)
 * P.S. I have the same issue with ion. There is no way to add a charge to the current template.  This seems fixable, though.
 * That didn't do it... Is this OK instead? It god rid of the whitespace for me. As for the formulae, I prefer free base whenever possible (drop the acid), but I don't think there is a consensus. Fvasconcellos 23:56, 24 February 2007 (UTC)
 * If it works, it works. I think I like the freebase too, but I'm going to continue ignoring them for now.  --Selket Talk 00:21, 25 February 2007 (UTC)

I've seen the results, and I personally don't really like them. The different colors makes it look very mickey mouse. Over at wikiproject chemistry, we discussed this and I think quite a few were not in favor of linking chemical formulae. Does it really help improve accessibility to wikify bromine and nitrogen? If you don't know what it is already, reading the article probably won't help your understanding of the drug. --Rifleman 82 06:02, 25 February 2007 (UTC)


 * I think it's useful for people who know some chemistry but not too much. I think a lot of the people who care about the structures of various drugs fall into this category.  Granted, it is unlikely that people are going to need the carbon link, but when I came across Gadodiamide (an MR contrast agent) for the first time I saw Gd in the formula and had no idea what that was.  I knew it wasn't gold, bit that was about it.  I ended up needing to google for a periodic table and then try to find Gd, then I went back to wikipedia to see what the deal was with Gadolinium.  I do agree that the colors can be a bit much, but I think it can serve as a useful key for the renderings (see Gadodiamide again) provided that we color them consistantly.  I have tried to use the same colors, but I can't speak for everyone.  --Selket Talk 06:41, 25 February 2007 (UTC)
 * In the chemistry articles we try not to colour/link chemical formulae. I personally don't like the elements being linked in a formula because it is distracting.  I think it is better to talk about the constituent elements in the text.  It is mostly C, H, N, which I expect the majority of people to know what that means, and for the stranger elements I find it more convenient to link them in the text.  "Mercuryoxide (HgO) is an oxide of mercury" (if you also put chemical formula into that sentence, people should have enough information).  For the drugs it is similar, most elements are quite general, and the gadolinium in Gadodiamide should be named and explained in the text anyway (I think it even should be linked in the intro).
 * About the colouring, it gives a very, as it was described, Mickey Mousey feeling to see the colours in the pictures and the formulae, although I do understand that for people outside chemistry it does clarify. Maybe the medicine should get a section with a description of the molecule, synthesis and (bio)chemical function (where I mean a description of why it does what it does, for as far known).  I think that explains more than colouring of the elements matching the picture.  --Dirk Beetstra T  C 11:27, 25 February 2007 (UTC)
 * I like the new format, because (1) it enforces rigor and consistency, which facilitates reuse and (2) it makes that line look different than the others, which makes it easier to orient the eye. But perhaps it would address the concern if the colors for the elements were less saturated. For example, instead of oxygen using "232 70 70" as it currently does, it could use something like "139 0 0". (A useful list of color codes is available at web colors.) --Arcadian 15:03, 25 February 2007 (UTC)

Unfortunately, there are several other templates for chemical elements. That said, the majority of pages written by members of chemistry wikiproject simply use super and subscripts. I have made my comments regarding the wikilinking of chemical formulae known already. To reiterate, it adds little or no value to the article to link to nitrogen or bromine, and if something unusual such as gadolinium is a part of the drug molecule, it should be discussed in text, per Beetstra.

SVG
Fvasconcellos, would you please share your method of making SVGs? I do agree they are superior, and that they should be used where possible. However, I'm not able to find a good way to export from ChemSketch to any svg editor, either through the clipboard or through EMF/WMFs.

That said, There are a lot of poor quality SVGs out there which have replaced better quality PNGs. Content is key, and not the format. --Rifleman 82 14:31, 25 February 2007 (UTC)
 * My method, despite (IMHO) good results, is not very efficient: I export as TIFF from ChemSketch at 600 dpi, import the file into Inkscape and trace the structure. I trace the structure manually, using the Path tool and angle constraints—this is very important, as the built-in automatic tracer gives low-quality results—and add atom labels as text. For 600 dpi TIFFs, I use the following Inkscape style settings:
 * 6 px line width for bonds
 * 82 pt Arial for atom labels
 * 65 pt Arial for subscript/superscript labels
 * Works for me :) If you have access to Adobe Illustrator, Ben Mills' cut-and-past method sounds far less troublesome—he outlines it on his Talk page. Fvasconcellos 15:29, 25 February 2007 (UTC)


 * Yes I have seen Ben's talk page prior. Like him, I'm not going to spend a bunch of money for Wikipedia.


 * I must say your work *is* excellent. But this method takes too much time for me. I feel that "good enough" is better than "perfect" later, and given my limited time, I'll stick with ChemSketch --> TIFF --> Irfanview --> PNG for now. It only takes a minute or so to draw a molecule and convert to PNG with this method. I think the results are wholly acceptable. --Rifleman 82 15:42, 25 February 2007 (UTC)

There is an alternative that is quicker and free but not so polished. BKchem is a simple freeware molecule editor. You can draw by hand or import from several import formats then export direct to SVG. There is a default 10px margin, which works well. For example, Dronedarone-2D.svg was created by entering the SMILES string from PubChem, rotating it and pressing Save. You can compare that with Fvasconcellos' version at Dronedarone.svg and Jfdwolff's PNG at Dronedarone.png. Oh, and if you want to know why there are three diagrams for one drug: Fvasconcellos and I both created versions on Commons and then discovered that Jfdwolff had beaten us to it. I really should get round to deleting mine, which certainly isn't the best of the three. For complex diagrams, like Sirolimus.svg, Fvasconcellos is the master! Colin°Talk 18:24, 25 February 2007 (UTC)
 * Thank you :) Fvasconcellos 15:33, 26 February 2007 (UTC)
 * Yeah, I tried that once, but didn't really like it. Perhaps I'm too accustomed to ChemSketch, but in the brief period I played with it, I couldn't really make it do what I needed to do/do nice structures with it. Still waiting for a way to copy directly into inkscape from ChemSketch through the clipboard. --Rifleman 82 18:49, 25 February 2007 (UTC)


 * To do it with ChemSketch, instead of exporting to TIFF, export to Windows Metafile (*.wmf). This can be imported into Inkscape as a vector object.  It will not include the text lables, so you will have to add them in Inkscape but it gives very nice results.  (See: Halometasone) --Selket Talk 16:52, 2 March 2007 (UTC)
 * I'm strangely unable to import WMFs to Inkscape. Wonder what could be wrong... Are you using an extension? Fvasconcellos 22:01, 3 March 2007 (UTC)
 * I'm using Inkscape 0.45, built Feb 5 2007 from Help->About on Windows XP.  I was able to do it with either File-Open or File->Import.   What version are you using?  --Selket Talk 22:44, 3 March 2007 (UTC)
 * 0.43, Nov 23 2005. Time for an update :) Fvasconcellos 22:55, 3 March 2007 (UTC)
 * OK—updated, and your method is officially approved! :) I still need to tweak the settings on both ChemSketch and Inkscape a bit, but, man, is that a time-saver. Thanks for sharing! Fvasconcellos 01:58, 4 March 2007 (UTC)

Bug with formulas using Cl atom! (and possibly other two lettered atoms)
I noticed a serious bug with this otherwise fine template. For some reason, it seems to change the place of Cl atom by putting it one off target in the formula.

Examples: Chloramphenicol formula is given as | C=11 | H=12 | N=2 | O=5 | Cl=2 but shown in DrugBox as C11H12N2Cl2O5

Clozapine formula is given as | C=18 | H=19 | Cl=1 | N=4 but shown in DrugBox as C18H19N4Cl

Ketamine formula is given as | C=13 | H=16 | Cl=1 | N=1 | O=1 but shown in DrugBox as C13H16NClO

Zopiclone formula is given as | C=17 | H=17 | Cl=1 | N=6 | O=3 but shown in DrugBox as C17H17N6ClO3

Sodium_thiopental formula is given as | C = 11 | H = 17 | N = 2 | Na = 1 | O = 2 | S = 1 but shown in DrugBox as C11H17N2O2SNa

Also:

Fluoxetine formula is given as | C=17|H=18|F=3|N=1|O=1 but shown in DrugBox as C17H18NF3O

I'm sure there are others as well. I don't know what's wrong with the last one, but the others seem to somehow originate from the number of letters (or so it seems).

In my opinion, this is a serious bug, because it changes the formula, making it impossible to use it as such, for instance, for Google searches. Hopefully somebody can easily fix this, as the DrugBox seems to be otherwise perfectly OK. —Preceding unsigned comment added by 82.181.234.205 (talk • contribs)
 * The formulas in the infoboxes are sorted into a standard order (Hill system), but you are right, the order seems to be off here (though not in the way you expect if I read your comment correctly). Formula's should be sorted C first, H second, all other elements in alphabetical order, in which case


 * | C=11 | H=12 | N=2 | O=5 | Cl=2 should display as C18H19Cl2N2O5


 * I'll have a look and resort them, if others agree. --Dirk Beetstra T  C 11:48, 27 February 2007 (UTC)


 * I see now... so the order of input is not even supposed to be the order shown in the box. It's a bit confusing, but if that's how it's supposed to work, then it's fine. I actually got confused when on some page (which I already forgot) the text showed the "usual" order, and the DrugBox showed another order which provided 0 Google hits. When I browsed the pages through quickly I made a wrong conclusion of that being somehow linked to two-lettered markings, which was because (by a random choise) I found that only two-lettered atoms were in the "wrong" places. When I found Fluoxetine, I thought maybe my original theory was not necessarily right, or maybe there were some other bug as well.
 * But if things are working as they have been planned to work, then everything is OK.


 * However I think this should be mentioned in Notes on use part. Now the example is actually showing that the order of input is also the order of output. Had there been some info about Hill arrangement, I wouldn't have suspected a bug. So a couple of lines to describe how it works would be a good help there. Just a suggestion...


 * Maybe there could be a way to mention this in the DrugBox also? Of course many databases are linked to the box, but just copy-pasting the formula might cause some problems... For instance, Zopiclone is in the DrugBox C17H17N6ClO3, but in PubChem it is C17H17ClN6O3. With PubChem this isn't a problem, because it is linked, but I hope everyone could see my point here. —The preceding unsigned comment was added by 82.181.234.205 (talk) 21:14, 27 February 2007 (UTC).


 * The order of entry of any of the parameters is not important (could place H=1 as 1st parameter, the drug name next and C=12 as the last parameter after pregancy details). I agree perhaps the notes should be ordered in the sequence that the various elements will be listed - just edit the /doc subpage :-) David Ruben Talk 20:44, 12 March 2007 (UTC)

Suggested updates for antibodies and recombinant proteins
There are a large number of drugs that are not small molecules. I would propose adapting this template to take a multiple choice "type" field that can account for this. For example, antibodies should probably not show up in Category:Chemical pages needing a structure drawing. I had previously proposed (but didn't get around to developing) a sequence field to replace IUPAC name on recombinant proteins. This is an example of the sort of thing that could be fixed with a "type" field. Any thoughts (objections)? --Selket Talk 02:53, 4 March 2007 (UTC)
 * How about a field such as ? I don't know about the code required, though. Fvasconcellos 02:55, 4 March 2007 (UTC)
 * I can code it, I just wanted to clear major changes with the community first. If I do re-code it, I will make sure it is reverse compatible so that nothing has to be done to the old articles. --Selket Talk 03:10, 4 March 2007 (UTC)
 * I support the general idea, but let's not rush into implementation, since this will have a wide impact, and I could see some problems handling large peptides. What specific fields did you have in mind? --Arcadian 04:15, 4 March 2007 (UTC)
 * For peptides: sequence, protein from which derived; for antibodies: mono- vs. poly-clonal, source organism, target protein. Replace Category:Chemical pages needing a structure drawing with Category:Proteins needing image or some such category.  Any other ideas?  --Selket Talk 07:18, 4 March 2007 (UTC)
 * Just wanted to point out, that there is a protein. I think that box can do that part of the trick.  --Dirk Beetstra T  C 11:04, 4 March 2007 (UTC)
 * Whatever. I have just been bold, and added a tag to protein, and did a recat on chembox new and drugbox.  They have now their own category, which is part of the master-category Category:Chemistry pages needing pictures.  It will take some time to (re-)populate, but this cleans it out quite a bit.  Hope this helps!  --Dirk Beetstra T  C 11:25, 4 March 2007 (UTC)
 * Just wanted to point out that WP:MCB have their own Image Request page, I think they should handle proteins and whatnot. Any thoughts? Recats did make things more manageable BTW, at least in my opinion. Fvasconcellos 14:03, 4 March 2007 (UTC)
 * I have been bold, and added a section to their image request with this category. I think this makes it indeed more manageable, though for the drugbox and chembox new the type of images requested is just the same.  I guess people from the projects can now find their own image-requests, while the top-level category does neatly connect the projects and gives people a quick way of finding the people who are active in picture drawing in the other projects.  --Dirk Beetstra T  C 14:26, 4 March 2007 (UTC)
 * Thanks, Dirk. We should also think about what to do about images for the antibodies.  Any ideas? --Selket Talk 03:43, 6 March 2007 (UTC)
 * There seems to be no objection to the general idea, Selket, so if you wanted to put those fields in, it would probably be okay. --Arcadian 19:55, 8 March 2007 (UTC)
 * Ok, its a major change, so I'm going to debug it pretty thoroughly in my user space first. It may be a day or two before I make the changes "go live".  --Selket Talk 20:00, 8 March 2007 (UTC)
 * crazy idea warning* Why don't we have a "MabBox" or "Drugbox monoclonal antibody"? We do have something like 200 monoclonal antibody articles. (I do realize this contradicts my previous suggestion of a "field".) Fvasconcellos 22:09, 8 March 2007 (UTC)
 * So, I added at drugbox-mab template per, Fvasconcellos' suggestion. It's not as elegant as a flag but much easier to code.  Feel free to add any other fields I didn't think of.  --Selket Talk 03:15, 12 March 2007 (UTC)
 * P.S. check out Bevacizumab for sample usage --Selket Talk 03:32, 12 March 2007 (UTC)


 * Hi, resonable idea, but are formula, melting points etc relevant for a protein ? Would an optional parameter in Template:Drugbox (that is normally not specified but defaults to say "drug") that if set to "mab" turns off various parameters not be an alternative method of coding ? (I think an alternative parameter name is mentioned higher up in this discussion). Otherwise it will be harder maintaining the 2 drug infoboxes. David Ruben Talk 20:40, 12 March 2007 (UTC)

Dose
could someone add a dose thing please, i have no idea how to do it —The preceding unsigned comment was added by The Right Honourable (talk • contribs).
 * WP:MEDMOS discourages the addition of such information to articles. How do you think it would be useful? Fvasconcellos 03:03, 17 March 2007 (UTC)

Multiple CAS numbers
I'm trying to clean up Methcathinone a bit by moving CAS numbers from the text to the drugbox. There are four listed and I'd like to have them appear with notes beside them like (racemic base), but still have the links work. CAS_supplemental seems not to do the links though, and CAS_number tries to put the comments in the links, even if the comment is surrounded by nowiki tags. Is there a simple way to do this, or do I need to just make the links manually?--Eloil 08:02, 9 April 2007 (UTC)


 * Use CAS_Number for the 1st number. Then additional numbers in CAS_supplemental using template to create links. See this edit to Methcathinone. David Ruben Talk 13:15, 9 April 2007 (UTC)

Links to FDA labels at DailyMed
Now that FDA labels are available at DailyMed, how about linking to these from the DrugBox? If you do this, you would pass the drug's generic name as the search term to DailyMed, then the user would select which preparation to view. Examples are: Badgettrg 20:35, 14 April 2007 (UTC)
 * http://dailymed.nlm.nih.gov/dailymed/search.cfm?startswith=sumatriptan
 * http://dailymed.nlm.nih.gov/dailymed/search.cfm?startswith=reserpine
 * We should compare this with the licence_EU and licence_US fields as discussed earlier at . Colin°Talk 20:41, 14 April 2007 (UTC)


 *  Does not this link provide essentially the same function as the licence_US parameter link (http://www.accessdata.fda.gov/scripts/cder/drugsatfda/index.cfm?fuseaction=Search.Search_Drug_Name) - i.e. to a datasheet. If so, then is one link system to be preferred (i.e. do we need change how the licence_US parameter is linked?), or does it vary drug by drug as to which gives the better information for a general readership (i.e. do we need an alterative parameter)? David Ruben Talk 20:45, 14 April 2007 (UTC)
 * Actually, last time I checked, DailyMed (which I've used as a reference for iohexol) didn't provide as many information sheets as AccessData. Fvasconcellos (t·c) 20:48, 14 April 2007 (UTC)


 * If so, then there is no need duplicate links for the same information and lets leave licence_US as is and no need additional link system :-) David Ruben Talk 20:54, 14 April 2007 (UTC)
 * I believe this is really so—just found the following disclaimer on the DailyMed website:


 * At the present time this Web site does not contain a complete listing of labels for approved prescription drugs.


 * Nice idea though. Drugs@FDA (AccessData) also does not contain labels for all Rx-only drugs available in the U.S., but there's no such disclaimer over there :) Fvasconcellos (t·c) 21:08, 14 April 2007 (UTC)

Revisit of this issue
I think we should reconsider this. Drugs@FDA is good, but for different purposes. Relying on Drugs@FDA for labeling information presents two problems
 * 1) The label "License data" is unlikely to suggest to anyone that this link will lead to a package insert. It sounds like it is going to lead so some legal stuff.
 * 2) Even if someone happens to take this link, frequently they still have to follow additional links at the FDA website to reach the labeling information for their drug.

I proposed the that DailyMed link be added and be labeled along the lines of "Package insert (US)" or maybe "Labeling information (US)". Admitted, this new source does not always have a label, but it can be an extended field and the editor only populate this field when the label is available.

Currently Wikipedia does not do a good job of linking the user to good, basic drug information. Badgettrg 17:38, 24 June 2007 (UTC)


 * DailyMed seems like a good source, but it isn't currently widely used, and I think we need to be conservative about adding fields to the infoboxes. I'd recommend proceeding in four steps. (1) Start adding links in the External Links section to a few pages. (2) Write a DailyMed article, to make it easier for people to understand the source. (3) If nobody complains after a month or so, then build a standalone template (like Template:Dailymed) and start using that, though still keeping the link at the bottom of the page. (4) If, after using it for another month or so in that format, then repropose it here, and if there are no major objections, then we could add it to the template. In addition to generating consensus, the experience we'd gain in using the stand-alone DailyMed template would help us figure out the right way to set up the parameters if/when it was added to the DrugBox.--Arcadian 21:42, 24 June 2007 (UTC)


 * Care needed not to duplicate info that this site uses that comes from FDA itself, and likewise not have duplicating parameters of Licence_US parameter and for this site. Also DailyMed lists for one specific brand of product, rather than the drug itself (no acetaminophen, but loads of "acetaminophen with codeine"). However Arcardian's suggestion for template seems sensible and I have created a Template:DailyMed (use of not just capitalised 1st letter is in keeping with Template:eMedicine, Template:WhoNamedIt etc). David Ruben Talk 23:38, 24 June 2007 (UTC)
 * As a single example of use, see this edit to Pregabalin article. David Ruben Talk 23:59, 24 June 2007 (UTC)
 * Looks good. --Arcadian 01:12, 25 June 2007 (UTC)
 * Indeed—a good chance for me to revisit my previous opinion :) Do you think there is a rationale for adding DailyMed to a few more pages as a trial, or should we start "spamming" it across the board? Fvasconcellos (t·c) 02:01, 25 June 2007 (UTC)
 * Anybody mind if I add this to the infobox? --Arcadian (talk) 02:32, 3 December 2008 (UTC)
 * I have added this field to the infobox. An example of it in use (and a demonstration of the field added) is visible here. --Arcadian (talk) 21:49, 19 January 2009 (UTC)

Given DailyMed links are far easier to get to information (i.e. fewer nested levels of pages) than the FDA site itself, and indeed seems to duplicate the product license details, two queries:
 * 1) If this patient-friendly link is specified, should licence_US parameter be disabled (i.e. surely only one of these two need be given),
 * 2) Should DailyMed link appear in chemical details section (as present) or move down to appear in the licencing/usage section along with licence_EU & licence_US (which further strengthens my 1st point, and follows logic of how I have added this to the documentation subpage  ? David Ruben Talk 01:05, 20 January 2009 (UTC)
 * I've got no objection to that. --Arcadian (talk) 03:54, 20 January 2009 (UTC)

Ok so recoded (and tested). In doing so I've made following choices:
 * as 2 options for US linking, we need show to which one (DailyMed vs FDA)
 * but still need to indicate that FDA & DailyMed both US sources
 * merely stating FDA equivalent of "EMEA" alone is not as obviously european without a similar EU prefix
 * Previously the external link was applied to the displayed text "US" or "EU", but as these now form part of source description we need find alternative. For consistancy across the 3 parameters, I chose "link" as the displayed text

Just DailyMedID

Just licence_EU

Just licence_US

Daily Med & licence_US specified

DailyMedID & licence_EU specified

David Ruben Talk 12:13, 20 January 2009 (UTC)

Links to Orange Book
Any interest in incorporating links to the FDA's orange book (http://www.fda.gov/cder/ob/) listings on particular drugs to the template? Remember 13:47, 4 June 2007 (UTC)
 * It might be interesting, although Drugs@FDA (already in the template, see above) probably provides information more relevant to the reader. It is an excellent source for checking generic availability and patent stuff, though. I'd like to hear more on this. Fvasconcellos (t·c) 13:56, 4 June 2007 (UTC)