smarts query

The following smarts will not return results in the PP chemaxon query component, but in IJC it does return results.

[CX3;$([R0][#6]),$([H1R0]):1](=[OX1])[F,Cl,Br,I:2]

PP chemaxon query component will return hits if each atom is explicit instead of having [F,Cl,Br,I:2]
[CX3;$([R0][#6]),$([H1R0]):1](=[OX1])[ClX1:2]
[CX3;$([R0][#6]),$([H1R0]):1](=[OX1])[BrX1:2]
[CX3;$([R0][#6]),$([H1R0]):1](=[OX1])[IX1:2]
[CX3;$([R0][#6]),$([H1R0]):1](=[OX1])[FX1:2]

Let me know, what information do I need to provide,

Hi,

Are you using the JChemSearch component on the same JChem table as with Instant JChem ?

Pleas let us know otherwise.

Please also let us know the settings (parameters) of the component (e.g. screenshot)

Please send us the version for both software.

In Pipeline Pilot you can find it at the top part of the output of  Component -> ChemAxon -> Utilities -> Show ChemAxon environment.

In IJC usually it's in the header or in Help -> About

Best regards,

Szilard

>>Are you using the JChemSearch component on the same JChem table as with Instant JChem ?

I used the same data set, imported the data into IJC and via jchem manager

>>Please also let us know the settings (parameters) of the component (e.g. screenshot)

ChemAxon For Pipeline Pilot System Information
  JChem version : 5.4.1.0
Component collection version : 1.8.1_j54
JVM: Sun Microsystems Inc.  Java HotSpot(TM) 64-Bit Server VM  1.6.0_01
Memory:  455.0 MB maximum  8.0 MB total  6.0 MB free

OS: amd64 Linux 2.6.18-92.1.18.el5
------------------------------------------------------------------

Thank you for the detailed information !

Forgot to ask ... could you also attach at least one hit structure from IJC ?

It would help a lot in the investigation.

If the structures are confidential you can also send them in e-mail to our technical support e-mail. (support@...)

.

see attachments

In the attachments you should have:

-  the sdf export of a subset, around 800 structures

-  the PP protocol

-  screenshot of the IJC query results [CX3;$([R0][#6]),$([H1R0]):1](=[OX1])[F,Cl,Br,I:2]

I did run the query against the same table, this are my results

smarts                                                                                 |        PP              |                     IJC

[CX3;$([R0][#6]),$([H1R0]):1](=[OX1])[F,Cl,Br,I:2]        |    0                      |                     42
[CX3;$([R0][#6]),$([H1R0]):1](=[OX1])[ClX1:2]             |    51                   |                     39
[CX3;$([R0][#6]),$([H1R0]):1](=[OX1])[BrX1:2]             |      2                    |                      2     
[CX3;$([R0][#6]),$([H1R0]):1](=[OX1])[IX1:2]                |     0                     |                     0
[CX3;$([R0][#6]),$([H1R0]):1](=[OX1])[FX1:2]               |     1                     |                     1

Hi,

Thank you for the report and all the detailed information !

I could reproduce the problem.

The problem was simpler than I suspected.

When there is a need to convert between CXN and PP molecules, the mol (CTAB) format is used as the best available choice.

Unfortunately this format does not handle SMARTS features well.

The conversion may also depend on the PP version, in my case with 8.5 I got 54 hits, not 0 (42 is correct).

 Suggestion:

- Use "ChemAxon MolConverter" to read in your SMARTS queries

- Specify a property name for output source, e.g. "mrvquery", choose mrv format 

- Specify the same property for the "Query Source Property" parameter of JChemSearch.

This way I got 42 hits for the first query, which is the correct number.

Later we will enable "JChemSearch" to read queries from a file directly as well, simplifying this.

Please let us know if this helps.

Best regards,

Szilard

thank you!