how to run compr with floatin point descriptors from cxcalc

I tried to run compr with the output of cxcalc. the two files look like this:








id Mass Polar surface area acceptorcount


1 439.328 91.63 4


2 379.132 37.66 6


3 457.344 107.37 3


4 390.262 97.36 7


... .... ... ... ...





I ran:





compr -z -m 3 -i firstfile secondfile > result





The error I get is Parameter -- fingerprint-size missing





Can anyone just give me an example for running compr with floating point descriptors, as I did not find any in the docu.





Thanks in advance!

Hi Markus,





I am substituting Miklós, who is the expert in these questions. It seems to me from the compr documentation, that compr input should be generated by generatemd, and not cxcalc:





http://www.chemaxon.com/jchem/doc/user/Compr.html#input width="90%" cellspacing="0" cellpadding="3" border="0" align="center"> Quote: Text input files can be created using the GenerateMD application. For example





generatemd c structures.smi -k CF -c cfp.xml -D -o fingerprints.txt See also the scalar section of the generatemd documentation:





http://www.chemaxon.com/jchem/doc/user/GenerateMD.html#output





Regards,


Szabolcs

O.K,


I just thought that it would be nice to have all the values of cxcalc to work with.


Another question:


If I want to retrieve the 40 most similar compounds of database1(L1), compared to each compount of database2(L2). How does compr autoscale the descriptor matrix?


do you calculate the mean of the each descriptor row within L2, or do I have to do that manually?





Thanks in advance

Hi Markus,





I am afraid we need to wait for Miklós about these questions, as I am not able to answer them. He will be back from holiday at the end of July.





Szabolcs

Hi Markus,





you also need "-f 0 " (zero, not letter O) to specify that no fingerprint is present in the input, then it should work as you expect.





There's no problem to use the output of cxcalc or any other application (including 3rd party ones) as long as it can generate a tabular text file in which each record occupies one line and values are separated by spaces.





Hope this helps,


Miklos