 |
|
| FriendFeed is the easiest way to share online. Learn more » |
Markus Sitzmann › Comments
so pretty to look at, but so simple to make (r + ggplot2) #rstats http://twitpic.com/85cddo
 |
Jan Wessnitzer, Egon Willighagen and Andrew Lang liked this
@Raj would you be able to share some code? I would love to see how these were made...
- Jan Wessnitzer
@Jan, sure. See https://gist.github.com/1596476
- Rajarshi Guha
sweet! I am wanting to visualize prior and posterior probability density functions and I think your code could be very useful indeed.... please do look out for some acknowledgment once I am successful ;)
- Jan Wessnitzer
Happy to be of help :)
- Rajarshi Guha
from iPhone
I use it from python via rpy2 - it is really nice
- Markus Sitzmann
I am either very confused about timezones or I'm getting tweets from @kaythaney from about ten minutes in the future...
Kubke, Dan Hagon and Andrew Lang liked this
Hmm, probably not enough time to ask her for any lottery numbers :-)
- Markus Sitzmann
Chemists: is there a short term to describe that a compound has no stereo-chemistry assigned? http://t.co/5qahAGJ vs http://t.co/N9cJ4TN
Egon Willighagen liked this
"stereochemically ambiguous", maybe?
- Rajarshi Guha
'mixture', I suppose.
- gilleain
either "unknown" stereochemistry or "unspecified" stereochemistry - depends on the case. "unspecified" is the 'worst' case - i.e. there is simply nothing at all about the stereochemistry of a compound. If a stereocenter is marked as "unknown", this already implies, somebody has already thought about it but the information is just not available.
- Markus Sitzmann
I am not aware of such terms yet. Rajarshi's suggestion is not complete, while I think the one buy Gilleain is incorrect. Let's make something up; it's important; without a name it does not exist, and it would be good if more databases would start to take this seriously. What a 'flat chemgraph' versus a 'steric chemgraph'? the chemgraph is just short for chemical graph, while the...
more...
- Egon Willighagen
Maybe it's Schrodinger's Compound (by analogy with the infamous cat). Its wavefunction has not yet collapsed, etc.
- Noel O'Boyle
I thought that meant racemic
- Andrew Lang
@Egon: http://old.iupac.org/publica... (page 1964 inparticular)
- Markus Sitzmann
"racemic" is always a mixture of two stereoisomers of know stereochemistry
- Markus Sitzmann
@Markus... yes, but 'unspecified' does not directly lead to a short term for such a compound... 'unspecified compound' does not do it, 'compound with unspecified stereochemistry' is not short anymore. If the search is just for the qualification, yes, then 'stereochemically unspecified' would do. And, as the material from Brecher says, it is of utmost importance that there is a clear indication if something is a mixture or not...
- Egon Willighagen
NCI database meeting has a great lineup (http://t.co/ZBY6FFO). Unfortunately, have to skip this meeting :(
tony yuan and Jean-Claude Bradley liked this
How come?
- Andrew Lang
need to fly to denver (ACS) that friday
- Rajarshi Guha
from iPhone
i'll miss it too -- hope it'll be a regular event!
- Geoffrey Hutchison
too bad we won't see you there - conflict with ACS is rough
- Jean-Claude Bradley
Sorry for that - we decided for this dates because of allowing people from Europe or the rest of the world for a combined travel to our meeting + ACS meeting
- Markus Sitzmann
no prblem - in fact I would've come if I were still in MD. But I'm moving to CT this week, so I got lazy..
- Rajarshi Guha
OC[C@@H](O)[C@@H](O)CO - http://apps.ideaconsult.net:8080/ambit2...
gilleain liked this
OB depiction now in ambit web service...Thanks Nina!
- Noel O'Boyle
Nice. Very nice ... but is it me or is that vertical H (at the end of the dashed wedge) slightly to the right? Text box bounds might be off.
- gilleain
Weird. An unrelated bug report on the cdk gives me this: http://apps.ideaconsult.net:8080/ambit2...
- gilleain
@gillean: Yes to both. (1) It's slightly to the right. (2) The end of the bond is truncated when the bond is stretched. There are a few other known problems - probably the most serious occurs for molecules with multiple rotatable bonds (too slow to layout). By and by we'll sort them out.
- Noel O'Boyle
However the bonds are cutoff nicely, something I wished we did properly. As for layout, well multiple rotatable bonds must be tricky. Dendrimers would be the worst, I suppose.
- gilleain
With multiple rotatable bonds, the code currently just brute forces all possible layouts. We should replace this with a stochastic search or a greedy search if the number of layouts is over a cutoff.
- Noel O'Boyle
Is it OK to say that I like the OB result better than the rest? :-) I do like the (R) and (S) designations with CACTVS, though.
- Geoffrey Hutchison
Cactus is most appealing to a bench chemist. If an undergrad wrote the Daylight version we would prob mark it wrong -- this is not a Fisher diagram.
- Orion Jankowski
from BuddyFeed
The (R) and (S) designations is something I switched on by using the appropriate CACTVS option during the image is generated by the Chemical Identifier Resolver (this is where the AMBIT services takes the images from) - but it isn't the CACTVS default behavior. However, I think every chemist should be trained by education to assign R/S, E/Z and undefined,unknown etc. to each structure she/he draws - would fix a lot of problems :-)
- Markus Sitzmann
Sean Ekins and I respond to Francis Collins letter "Reengineering Translational Science": http://www.chemconnector.com/2011.... Sci Trans Med didn't respond
Steve Koch, Jean-Claude Bradley and Andrew Lang liked this
Mhh, I can see it with a release day of yesterday
- Markus Sitzmann
I will never be able to blend fruits again :( - Imgur - http://imgur.com/I1lXn
Mickey Kosloff and Egon Willighagen liked this
Indeed!
- Markus Sitzmann
The alpha and omega of SMILES strings Part II - http://baoilleach.blogspot.com/2011...
 |
Rajarshi Guha and Jean-Claude Bradley liked this
Noel is there a way to represent correlated uncertainty in chirality? When there is only one chiral center, not specifying the chirality is a way to refer to a racemic mixture. But when there are two chiral centers is there a way to represent a racemic mixture of R,R and S,S (and no R,S and S,R) with only one SMILES string? This has been coming up frequently in the melting point literature.
- Jean-Claude Bradley
Chirality is specifed on a per atom basis, and so you could say that a particular atom is chiral, only you're not sure of the absolute configuration: C[C@?](Br)(Cl)I. It sounds like you are talking about a mixture of two different molecules. SMILES has a dot disconnected notation for unconnected components in a 1.1 mixture, so you would have MoleculeA.MoleculeB where A and B have opposite stereo specified.
- Noel O'Boyle
So a question mark is an actual SMILES character?
- Jean-Claude Bradley
Uh, actually, maybe not. Just been looking back through my emails and at opensmiles.org, and it seems that that was just in the devel code.
- Noel O'Boyle
OK. I guess the closest we can come is to put the two SMILES connected by a dot.
- Jean-Claude Bradley
I regard it as dangerous to use SMILES linked by dots for the representation of racemic mixtures - for the simple reason that even simple property calculations like molecular weight have a wrong result unless you mark the dot-linked SMILES explicitly as racemic mixture in some way (but if you do this you have to add something to the SMILES outside the SMILES anyway).
- Markus Sitzmann
Markus - you are right linking enantiomer pairs with a dot isn't exactly the right concept since the molecular weight would be of the entire "complex". The dot makes sense for salts.
- Jean-Claude Bradley
my blog traffic from search engines is from Google. Period. No trace of Bing. How is that for others?
Top 3 search engines: Google 95.1%, Yahoo 2.1%, Bing 1.9%
- Andrew Lang
I can confirm the numbers from Andy - they are almost identical for cactus.nci.nih.gov. Searches from "Google Image" generate more traffic than entire Bing. Also interesting: if the share of traffic using Firefox 5 (yes five) continues to grow like it did yesterday it will overtake IE9 tonight or tomorrow.
- Markus Sitzmann
so, my #ChemSpider submission is curated, approved, and published http://www.chemspider.com/Chemica... but where is that Open Data button?
Andrew Lang and Jean-Claude Bradley liked this
There is an Open Data box to check when you upload a spectrum
- Jean-Claude Bradley
Yes, and that checkbox is also there when you submit a new structure...
- Egon Willighagen
I have already contacted Antony about this... so, to be continued (in due time)...
- Egon Willighagen
Egon, is your structure correct? Didn't you mean http://www.chemspider.com/527879 (one of your rings is not aromatic)
- Markus Sitzmann
Will check shortly... ChemSpider is down...
- Egon Willighagen
Ouch... Markus, you are right... :(
- Egon Willighagen
I requested the paper to be linked from the structure I added, and added the DOI to 527879... peer review++
- Egon Willighagen
yeah, it also can be used for detection of aromaticity :)
- Egon Willighagen
Markus, is the lower left corner using your service?
- Egon Willighagen
As much as I can see from our log files: yes.
- Markus Sitzmann
gChem: Convert Names and CAS Numbers to Chemical Structures in Google Spreadsheets [Video] - http://metamolecular.com/blog...
joergkurtwegner, Jean-Claude Bradley, Jan Jensen and 3 other people liked this
That's awesome!
- Andrew Lang
Love it - what else should I say :-)
- Markus Sitzmann
Nicely done video too.
- Noel O'Boyle
Github Has Surpassed Sourceforge and Google Code in Popularity - http://www.readwriteweb.com/hack...
Deepak Singh and tony yuan liked this
I am not surprised - particularly I still don't understand what Sourceforge intended with their last web page redesign - horrible.
- Markus Sitzmann
Um, based on the number of commits? This is a comparison of git vs svn commit frequency, rather than popularity of platforms.
- Noel O'Boyle
Are there any hosts/websites/whatever which are IP6 ready?
- Markus Sitzmann
WordPress Discontinues Support for Internet Explorer 6 - http://www.readwriteweb.com/hack...
Me, too - if that matters :-)
- Markus Sitzmann
RT @drjonboyg: Things that make me go hmm: reading a claim that "Over 83 experiments have yielded the same results." Over 83? You mean 84? #sciwriterfail
Could be a way of hedging. The author may not know of all or new ones may have appeared recently.
- Yo. Shark Dog.
Maybe it were 83.5 :-)
- Markus Sitzmann
(Almost) Translate the InChI code into JavaScript Part II - http://baoilleach.blogspot.com/2011...
tony yuan, Markus Sitzmann and Egon Willighagen liked this
I am really curious what your end result will be :-)
- Markus Sitzmann
just discovered angry birds. this can only end badly :(
Egon Willighagen and Josef Scheiber liked this
Rajarshi, please don't post things like that again - I need to work :-)
- Markus Sitzmann
jStat : a JavaScript statistical library - http://www.jstat.org/
Rajarshi Guha liked this
One day we will do everything in the browser :-)
- Markus Sitzmann
Very useful!
- tony yuan
IE10 Platform Preview and CSS Features for Adaptive Layouts - http://blogs.msdn.com/b...
IE 10 - already!
- Rich Apodaca
If I look at the log files of our web server, they have a reason - IE9 is on par with Chrome 11 (yes eleven) - Firefox 4.0 is ten-fold ahead
- Markus Sitzmann
Out of interest, can someone from the #PistoiaAlliance tell me whether the website was all built with #Wordpress? Very nice! #BioIT11
I doubt it is WordPress - it looks like Joomla
- Markus Sitzmann
Perhaps the most bizarre and psychedelic page on solubility - just scroll through the whole thing and enjoy http://overeem.dk/tailori...
woah!
- Rajarshi Guha
Probably it looks perfect if viewed in Word 7.0
- Markus Sitzmann
another similar page - I don't know what the motivation is to create these but they are entertaining http://eye9d.com/footpri...
- Jean-Claude Bradley
RT @hleman: Web Sites Should Not Be Replaced by Facebook Pages http://blog.sysomos.com/2011... I wish everyone in the world would read this--important! Pls retweet.
Well, my bet is that Facebook will be the "GeoCities" of 2020
- Markus Sitzmann
Can't quite describe my incredulity. A game played entirely in your URL address bar - http://t.co/ZQLooPA - http://twitter.com/michael...
Egon Willighagen, Michael Wolfle, Orion Jankowski and 3 other people liked this
can we use this for science :-)
- Markus Sitzmann
Brilliant!
- Egon Willighagen
Markus: "move around the O in the URL bar, until you hit a SMILES which cures pancreatic cancer"?
- Egon Willighagen
Damn...I upgrade to the latest IE version and now cannot copy and paste text off of a web page? WHat the...What's behind that new function?
For copy&paste you need to install Silverlight ... no, I don't know - I am joking :-)
- Markus Sitzmann
The failed roundtripping of InChI to structure to InChI corresponds to a subset of 18,359 compounds out of 38,966,897 PubChem Compounds
Did you redo it or is this a repost from the numbers here: http://sourceforge.net/mailarc...
- Markus Sitzmann
good morning cold Uppsala! I'm fine with -23.4 celcius (normal here), but not on my bicycle... :(
rowlikeagirl, Claudia Koltzenburg and Shanthi Bradley liked this
Wow, DC had -12C last night and fresh snow - but tomorrow spring will be back
- Markus Sitzmann
in early December in Novosibirsk with some degrees below yours I saw a cyclist with a helmet plus nosewarmer, that really made sense to me :-)
- Claudia Koltzenburg
It's only around freezing here. Pretty warm winter so far in Berlin.
- Björn Brembs
You folks are telling me I should have packed more jumpers aren't you... :-)
- Cameron Neylon
It was -18C here in Warsaw when I was going to work yesterday (35 minutes walk).
- Pawel Szczesny
from iPhone
gChem: Easily Convert Names and CAS Numbers to Chemical Structures in Google Spreadsheets | Signals - http://blog.metamolecular.com/article...
joergkurtwegner and Egon Willighagen liked this
Looks nice, Rich - need this for our chemical inventory. What's the hit rate?
- Matthew Todd
from Bookmarklet
@Mat, hit rate was found to be ~70% for one sample set. See poster in http://depth-first.com/article...
- Rich Apodaca
Does it do InChI 2 InChI, or InChI to (ChemSpider|PubChem) ID?
- Egon Willighagen
InChI 2 InChI -> to remove/add layers
- Egon Willighagen
Since it gets its data from the Chemical Identifier Resolver: ChemSpider IDs: yes, PubChem SID: yes (but a little bit outdated), PubChem CID: no (is on my ToDo-List), InChI Options: (is on my ToDo-List)
- Markus Sitzmann
Different Approaches to Storing Structures and InChIs on ChemSpider - http://www.chemconnector.com/chemuni...
Orion Jankowski, joergkurtwegner, Jean-Claude Bradley and Egon Willighagen liked this
Speaking of tautomers, why does ChemSpider often pick the minor contributor for amides or carbamates when I deposit them on CS?
- Jean-Claude Bradley
Hmmm...are you saying that when you deposit a particular tautomeric form that the deposition system then changes it to another tautomeric form that ends up being deposited? If so, can you point me to a particular example pls...
- Antony Williams
Guys...what are your thoughts about the multiple tautomers? The observation of multiple tautomers in ChemDB and emolecules, for example, really leads to bloated database statistics. What is your preference? ONE structure with links to multiple tautomers or store all tautomers separately?
- Antony Williams
Depends on how you want to present calculated properties... if one entry for all tautomers, then you need to give all 2D depictions, and 3D models, and all calculated properties, without making the user dizzy... what about 42 warfarin tautomers? on one page?
- Egon Willighagen
OK, how about this idea... it does require some serious hacking, but would be clear to the user, and solve your redundancy worries: have one entry, show 1 2D diagram (and matching 3D), but allow the user to move left and right, iterating through the 2D diagrams of the various tautomers, while updating the stuff that depends on it, like 3D model (with hydrogen) as well as calculated...
more...
- Egon Willighagen
My thoughts are as follows. ONE tautomer shown, others linked. Properties calculated for the ONE tautomer but other tautomers can be passed out to external predictors such as Chemicalize. The overhead to host all predicted data for all tautomers compared with the use that such a collection of data will receive is too large. It's the difference between what's possible compared with...
more...
- Antony Williams
But then at least be clear which tautomer is depicted (InChI with fixed-hydrogen layer) and also a clear indication that calculated properties, like ACD/logP, are for that tautomer.
- Egon Willighagen
Well, I would say the best way is: keep any original tautomer (as submitted by the source) and linked it to your canonical tautomer (whatever definition you have for this). Usually very minor tautomeric forms a never reported, so for a structure always only a few tautomers are interesting (and might be even separate compounds under certain conditions which are worthwhile to be presented...
more...
- Markus Sitzmann
Egon...yes, the InChI with fixed layer and PhysChem properties will be for the tautomer depicted
- Antony Williams
Markus...no apologies necessary re. the statement about the PubChem deposition. We are aware of some of the issues with the original deposition and have been in communications with Evan Bolton about that. Will you be at ACS???
- Antony Williams
So far no - except there will be some InChI meetings (but I haven't heard anything yet). I hope I get ICCS 9 funded
- Markus Sitzmann
Tony - draw CH3NHCOCH3 (usual tautomer) in ChemSketch and I get this InChI InChI=1/C3H7NO/c1-3(5)4-2/h1-2H3,(H,4,5) Hit deposit structure on ChemSpider and convert that InChI to image and the double bond is between C and N - the minor tautomer
- Jean-Claude Bradley
JC, might that because the other tautomer is already in the database? Have you tried a structure too of which there is no tautomer in the DB yet?
- Egon Willighagen
I just went through the process you outlined JC and I found the expected tautomer here: http://www.chemspider.com/Chemica...
- Antony Williams
I would say the minor tautomer is regarded as the canonical tautomer by InChI (if no fixed H layer is specified)
- Markus Sitzmann
Tony - I'm not talking about finding an existing structure - it only happens when I start to deposit a new structure. I would guess it is an issue with the software that converts the InChI to a structure during the deposition process.
- Jean-Claude Bradley
Egon - it happens with most Ugi products I deposit in ChemSpider (not previously there) - that is how I noticed this problem.
- Jean-Claude Bradley
That should be checked indeed...
- Egon Willighagen
JC...please work with me for your next deposition ...I'll need the molfile and I'll walk through the process. It could be that we need to introduce a /fixedH flag at deposition to retain the tautomer "as drawn"
- Antony Williams
I am presently working on a list of updated features so send me something in the next couple of days if possible. Thx
- Antony Williams
Actually importing the same InChI generated from ChemSketch back into ChemSketch also generates the minor tautomer! SMILES doesn't but we have had too many issues with SMILES (generated from ChemSketch) for depositing structures. The mol file re-imports into ChemSketch correctly too.
- Jean-Claude Bradley
Well, it is definitely the InChI library which does that - it occurs quite frequently if the fixed hydrogen layer is not present.
- Markus Sitzmann
Confirmed...this is an InChI library issue...not under our control unfortunately.
- Antony Williams
Hahaha, those knowing 'cheminformatics kernels' are surely not surprised, and why do I have this deja-vu feeling? http://ff.im/x9sJv IMHO is the current InChI code too weak in dealing with all the cheminformatics challenges we have discussed so many time e.g. on opensmiles.org . Whatever canonical form is produced I am in favor of seeing a more advanced cheminformatics kernel being...
more...
- joergkurtwegner
I'm a little bit perplexed by this dialogue. It seems to me that this isn't a deficiency of any particular representation, but rather a more insidious problem of "Guess what I mean." For any system to be useful (regardless of the final storage/rendered format) it will need to be robust to an underspecified input (or even incorrectly specified input), and rectify it to what you think an...
more...
- Orion Jankowski
BTW, there have been lawsuits over prior claims to enol tautomers that slipped into a Markush despite being chemically implausible as an isolated substance. The argument goes: you claimed the ketone, but my Markush covered the enol (by mistake) five years prior. Therefore, I own the compound. Depending on the ketone, the enol can be a very major microspecies (sometimes the only one). So...
more...
- Orion Jankowski
InChI is NOT the problem...it is simply a representation. Despite it's utility, and it is useful, it has many limitations of course. And Tautomers are the bane of most structure databases in terms of completeness of representation and challenges with search coverage
- Antony Williams
I also definitely don't see this as a limitation of InChI - it is a canonical representation of the structure by the rules and scope of InChI. That InChI chooses a structure that most people regard as the minor tautomer has no particular meaning - the important part is that any given tautomer of the original structure can be converted into the canonical tautomer by the algorithm. If a...
more...
- Markus Sitzmann
Drugs, Racemic Mixtures, Data Curation and Tautomers - http://www.chemconnector.com/chemuni...
You have the same problem for thalidomide (and the problems created by this compound are well known ...): http://cactus.nci.nih.gov/chemica...
- Markus Sitzmann
Other ways to read this feed:
