Feature request: 2000 chemical terms at your fingertip

User 677b9c22ff

16-11-2006 04:37:59

Hi,


I like the little drop down window, which opens in the chemical terms filter and has currently only:


Bioavailability


Ghose filter


Lead likeness


Lipinski rule of 5


Muegge filter


Veber filter





I am thinking about a really nifty feature which provides all terms from cxcalc, evaluator and all other possible ChemAxon and user resources.


http://www.chemaxon.com/jchem/doc/user/ChemicalTerms.html


It could even contain fingerprint features, like these which contain over 1000 fingerprints calculated from smarts/smiles:


ftp://ftp.ncbi.nlm.nih.gov/pubchem/specifications/pubchem_fingerprints.txt





The whole feature implemented as XML like in the TOOLS-Template-Manager would give a helpful resource to all users and could contain all the helpful material ChemAxon assembled over the years and which is sometimes hidden in the DOCs. These terms or features carefully sorted in XML-subdirectories and divided in default and user would be really nice. Because over the years ChemAxon added many additional properties which can be calculated and which are sometimes lost in space. Each user could add their own needed terms.





This feature list would also be helpful provided in/during


1) the "query window"


2) the "add columns window"


3) the "add 1000 columns with these selected features" window





I must admit the combined drop down menu with the originally 5 combined terms inspired this idea and the GUI and such handy tools can make the whole Instant-JChem a really powerful but easy to use tool.





Tobias

ChemAxon fa971619eb

16-11-2006 08:40:37

Two changes that will be happening at some stage should provide most of what you want.





Firstly the the chemical terms editor will provide code completion capabilitites, so that when you type in Ctrl-space you get a popup list of available terms, or if you have already typed the first few characters then you get the list of possible completions (you get descriptions of the terms in the popup list as well). This will make typing your own terms easier.





Secondly, we will be making the list of favourites user-editable, so that you can aded your own definitions to it.

User 677b9c22ff

17-11-2006 06:30:09

Hi Tim,


thats great, code completion is good. I like a tree-style selection also, because otherwise you dont know what kind of stuff you can select, but lets see first....


Tobias

User 677b9c22ff

12-01-2007 04:39:53

Hi,


I am not sure but, I remember Tim wrote that I can only define


the terms I want to calculate during DB import by adding them one after one.


Now it would be easy to code soemething in JAVA (I guess ;-)


so that one can calculate/import all terms at once during import, without typing them. See the table at end.





There are some atom based terms, but at least all the obvoius descriptors


would be nice, this could also solve the problem that constantly new descriptors or atom properties are added by ChemAxon, hence such a XML template could be updated with the Automatic Update Center too (a feature I really like a lot).





I dont know if it would make sense to add other features which require


the input of an additional value like pKa('pKa', '1') or such descriptors


which require the input of an atom number like stericHindrance(1) which would require an automatic drill-down analysis, hence when the molecule


has 14 carbon atoms it will provide a complete and automatic coverage


for atom 1..14 - ok maybe thats to complicated for Instant-JChem.






Tobias








Such an XML termplate table could/should include (+future extensions):


1 atomcount


2 composition


3 dotdisconnectedformula


4 elemanal


5 exactmass


6 formula


7 icomposition


8 iformula


9 mass


10 atompol


11 averagepol


12 avgpol


13 axxpol


14 ayypol


15 azzpol


16 charge


17 ioncharge


18 molpol


19 oen


20 orbitalelectronegativity


21 pol


22 polarizability


23 tpol


24 tpolarizability


25 conformers


26 hasvalidconformer


27 leconformer


28 moldyn


29 aliphaticatom


30 aliphaticatomcount


31 aliphaticbondcount


32 aliphaticringcount


33 angle


34 aromaticatom


35 aromaticatomcount


36 aromaticbondcount


37 aromaticringcount


38 asymmetricatom


39 asymmetricatomcount


40 balabanindex


41 bondcount


42 bondtype


43 carboaromaticringcount


44 carboringcount


45 chainatom


46 chainatomcount


47 chainbond


48 chainbondcount


49 chiralcenter


50 chiralcentercount


51 connected


52 connectedgraph


53 cyclomaticnumber


54 dihedral


55 distance


56 distancedegree


57 dreidingenergy


58 eccentricity


59 fusedaliphaticringcount


60 fusedaromaticringcount


61 fusedringcount


62 hararyindex


63 heteroaromaticringcount


64 heteroringcount


65 hindrance


66 hyperwienerindex


67 largestatomringsize


68 largestringsize


69 msa


70 plattindex


71 psa


72 randicindex


73 ringatom


74 ringatomcount


75 ringbond


76 ringbondcount


77 ringcount


78 ringcountofatom


79 rotatablebond


80 rotatablebondcount


81 shortestpath


82 smallestatomringsize


83 smallestringsize


84 stericeffectindex


85 szegedindex


86 topanal


87 wienerindex


88 wienerpolarity


89 canonicalresonant


90 canonicaltautomer


91 doublebondstereoisomers


92 resonantcount


93 resonants


94 stereoisomers


95 tautomercount


96 tautomers


97 tetrahedralstereoisomers


98 logd


99 logp


100 majorms


101 msdistr


102 pi


103 pka


104 acc


105 acceptor


106 acceptorcount


107 accsitecount


108 don


109 donor


110 donorcount


111 donsitecount


112 energy


113 hbda


114 huckel


115 msacc


116 msdon


117 order


118 pichargedensity


119 pienergy


120 refractivity


121 totalchargedensity

ChemAxon fa971619eb

12-01-2007 08:24:45

You mean you want to have all of those chem terms calculated for your structure tables? Really!


If this is the case then:


1. The import will be very slow as all the chem terms are being calculated


2. Some DBs will have problems as there are limits to the number of columns that can be in a table (32 for Oracle if a remember correctly, though this may be version dependent).





One thing that might be worth considering is defining your table with all the chem terms columns first and then importing the structures into it. That way adding the chem terms columns will be very fast as the table contains no strucutures, and the import will be slow as the calculations are being done as you import each structure. But his way you do not have to wait for each chem terms column to complete before you add the next, so the process should be easier.





In IJC 2.0 we have queing of adding columns, so that you can add a new one whilst a chem terms calculation is running. The addition get queued and starts when the previous one finishes.


We also plan to allow multiple columns to be added and calculated at once, but this is not yet implemented and may take some time to be ready.





Tim

User 677b9c22ff

12-01-2007 18:56:33

Hi Tim,
tdudgeon wrote:
You mean you want to have all of those chem terms calculated for your structure tables? Really!


If this is the case then:


1. The import will be very slow as all the chem terms are being calculated


I am a fan of speed :-) I still hope JChem is getting faster and faster from day to day. Also I can use an 8 core Opteron, so I dont really care how long it will take.





And I am interested in using Instant-JChem. For all other stuff I can probably use Dragon and some of our Opteron clusters.
tdudgeon wrote:



2. Some DBs will have problems as there are limits to the number of columns that can be in a table (32 for Oracle if a remember correctly, though this may be version dependent).


I mean for Instant-JChem, not for Oracle or MySQL....





[Oracle 10g Release: The absolute maximum number of columns in a table is 1000. However, when you create an object table or a relational table with columns of object, nested table, varray, or REF type, Oracle Database maps the columns of the user-defined types to relational columns, in effect creating hidden columns that count toward the 1000-column limit. ORacle.com


]
tdudgeon wrote:



In IJC 2.0 we have queing of adding columns, so that you can add a new one whilst a chem terms calculation is running. The addition get queued and starts when the previous one finishes.


We also plan to allow multiple columns to be added and calculated at once, but this is not yet implemented and may take some time to be ready.


Tim
Ok, my problem is, doing this one by one for 120-200 terms


will take 10 seconds each plus hundreds of mouse clicks.


Why not provide a template like in MVIEW the


Topology Analysis Options where all terms are clicked


already and then its scheduled and can calculate.





Maybe your implementation already looks like that,


and I am just talking in circles .-)


Tobias