SMARTS bug?

Hi all,





According to JChem documentation it should support v atom primitive for atom valence. Running the query

Code:

evaluate -e 'matchCount("[!#15v5]")' t1.smiles

where t1.smiles contains

Code:

FC(Cl)=CCl

returns 0





if I run the same command but without v atomic primitive the result is ok

Code:

evaluate -e 'matchCount("[!#15]")' t1.smiles

returns 5





Oleg.

Hi,





Because none of the atoms in your smiles "FC(Cl)=CCl" have valence of 5, I believe that both results are correct. What would you like to describe with your SMARTS expression?





Regards,


Szabolcs

Hi,





I am screening a library of compounds for SMARTS pattern

Code:

[!$([#6,F,Cl,Br,I,o,$([#8](-[#6a])-[#6a]),s,nX3,$([#7]C=O),$([#7]-[#6a]),#7v5,#15v5,#16,*+1,*+2,*+3])]

the SMARTS pattern

Code:

[!#15v5]

is part of the above pattern, anyway according to the Daylight SMARTS depiction

Code:

http://daylight.com/daycgi_tutorials/depictmatch.cgi

the structure in t1.smiles shouldn't be a hit while doing the search with JChem


gives 1 hit. Is there a bug in JChem SMARTS parsing/depiction?





Oleg.

OK, I spotted it. There is a H match. If you click "Enable explicit-H SMARTS" on depictmatch, it will also report a hit.





You can use this modified SMARTS if you want to exclude Hydrogens:

Quote:

$ evaluate -e 'matchCount("[!$([#6,F,Cl,Br,I,o,$([#8](-[#6a])-[#6a]),s,nX3,$([#7]C=O),$([#7]-[#6a]),#7v5,#15v5,#16,*+1,*+2,*+3])!#1]")' "FC(Cl)=CCl" 0

The important bit is "!#1" at the end of the SMARTS expression.





Best regards, Szabolcs

Hi Szabolcs,





Thanks a lot for quick replay, it does work now - I overlook the fact that JChem treats implicit and explicit hydrogens in the same way.





thank you for your support!





Oleg.