[Dock-fans] SPHGEN error

Eric Pettersen pett at cgl.ucsf.edu
Thu May 28 17:18:46 PDT 2009


Hi Michael,
	Here's the pertinent C code from DMS:

      			fprintf(fp, "%3s %4s %4.4s%8.3f %8.3f %8.3f %3s %6.3f",
                                 ap->restype, ap->resseq, ap->atname,
                                 pp->coord[0], pp->coord[1], pp- 
 >coord[2],
                                 typep, sp->area);
                         if (have_normal) {
                                 fprintf(fp, " %6.3f %6.3f %6.3f", np- 
 >coord[0],
                                         np->coord[1], np->coord[2]);
                                 np++;
                         }

	Matching those printfs against the formats you listed it seems that  
the "chain id" one would work if the first f7.3 were instead an f8.3.   
I'll send you Pasha's DMS file separately.

--Eric

On May 28, 2009, at 5:05 PM, Michael Mysinger wrote:

>
> Well, looking quickly, it first tries this format without the chain  
> id:
> format(a3,i5,1x,a4,1x,2f8.3,f9.3,1x,a3,7x,3f7.3)
>
> then this one with the chain id:
> format(a3,i4,a1,1x,a4,1x,f7.3,2f9.3,1x,a3,7x,3f7.3)
>
> and finally the one below, and fails on all. Indeed, there is a  
> suspicious inconsistency between the first 2 format statements near  
> the first floating point numbers.
>
> Pasha, would it be possible to send the original .ms file that  
> generated this error?
>
> Eric, do you know which of "1x,2f8.3,f9.3,1x" or "1x,f7.3,2f9.3,1x"  
> is more accurate?
>
> Cheers,
> Michael Mysinger
> UCSF DOCK Team
>
> Eric Pettersen wrote:
>> On May 27, 2009, at 10:30 PM, pasha shak wrote:
>>
>>> apparent state: internal I/O last format:  
>>> (A3,I3,A1,2X,A4,1X,F7.3,2F9.3,1X,A3,7X,3F7.3)
>>
>> Interesting.  This is not the format that DMS puts out (neither the  
>> Chimera version nor the compiled version).  I believe the format  
>> statement above will skate by if there is no chain ID.  If there is  
>> one, it will blow up.
>>
>> Perhaps the format worked with the ancient ms program, but I don't  
>> have access to that anymore.  At any rate, sphgen needs fixing if  
>> it's going to work with DMS.
>>
>> --Eric
>>
>>                        Eric Pettersen
>>                        UCSF Computer Graphics Lab
>>                        http://www.cgl.ucsf.edu
>>
>>
>



More information about the Dock-fans mailing list