ECFP as descriptor

Hi,

How to generate fixed length ECFP??, I tried,

generatemd c input.sdf -k ECFP -2 -o out.sdf -S

Which generates ECFP as List of integer identifiers.

Please help.

-nilesh

Nilesh,

according to the documentation from

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

   generatemd c input.smiles -k ECFP -c ecfp_config.xml -2

 

Because there is no ecfp_config.xml to be found on the internets and archives, you will go to your

installation \\jchem\examples\config\ecfp.xml and use this file as example.

or you can see it here:

http://www.chemaxon.com/jchem/examples/config/ecfp.xml

 

According to the documentation you need to change the three main parameters:

"The three main parameters of ECFPs are maximum diameter, fingerprint length, and identifier counts."

 

So you would replace the original files

"<Parameters Length="1024" Diameter="4" Counts="no"/>"

with

"<Parameters Length="512" Diameter="2" Counts="Yes"/>"

 

Then you run (assuming you use C6H6.smi as input file):

generatemd c C6H6.smi -k ECFP -c ecfp_config.xml -2

The output would be something like (right side snip):

Z:\>generatemd c C6H6.smi -k ECFP -c ecfp_config.xml -2
00000000|00000000|00000000|00000000|00000000|00000000|00000000|0
00000|00000000|00000000|00000000|00000000|00000000|00000000|0000
00|00000000|00000000|00000000|00000000|00000000|00000000|0000000
00000000|00000000|00000000|00000000|00000000|00000000|00000000|0
00000|00000000|00000000|00000000|00000000|00000000|00000000|0001
00|00000000|00000000|00000000|00000000|00000000|00000000|0000000
00000000|00000000|00000000|00000000|00000000|00000000|00000000|0
00000|00000000|00000000|00000000|00000000|00000000|00000000|0001
00|00000000|00000000|00000000|00000000|00000000|00000000|0000000
00000000|00000000|01000000|00000000|00000000|00000000|00000000|0
00000|00000000|00010000|00000000|00000000|00000000|00000000|0001
00|00000000|00000000|00000000|00000000|00000000|00000000|0000000
00000000|00000000|00000000|00000000|00000000|00000000|00000000|0
00000|00000000|00010000|00000000|00000000|00000000|00000000|0001
00|00000000|00000000|00000000|00000000|00000000|00000000|0000000

I attached the files for your convenience, hope this helps, have a nice weekend.

Tobias

Nilesh,

Did Tobias' answer help? Or do you have further questions?

 

Tobias,

many thanks for the great user support.

 

Regards

Miklos

mvargyas wrote:

Nilesh, Did Tobias' answer help? Or do you have further questions?   Tobias, many thanks for the great user support.   Regards Miklos

Thanks a lot, Tobias and Miklos. Answer was helpful.

-Nilesh

As a quick followup to this post:

Is it possible to output this binary formatted fingerprint to an SDF?  I tried 

>generatemd c sample_1000_molecules.sdf -k ECFP -c ecfp.xml -2 -S -o test2.sdf

and got the following:

Binary format is not allowed for SDfile output. Option -2 ignored.

Thank you for any help you can provide.

 

Dennis

Hi Dennis,

That's correct, binary string representation is not supported in SDF. Isn't decimal feasible? That can be reinterpreted as binary in your application. 

Miklos

I having trouble in generating the ECFPs its says BAD XML file ,parameter type and parameter configuration file not matching

I am using windows 7 .

Hi,

Can you send us the command line you tried? Without that it's not possible to figure out what went wrong.

 

Thanks

Miklos