FBEAUT error

Submitted by gongpu zhao on
Forums
Frealign

Has anybody met such a problem?
I use frealign to process a helical tube. It works perfectly for binned data (332X332 pixels). Once I start processing unbinned data (664X664), the alignment works fine, but the reconstruction did not generate any density in the map.The program ends normally, did not give any error information. I later found out this was due to the FBEAUT function. If I turn FBEAUT off, the reconstruction is fine.

niko Wed, 2011-01-05 07:49

I have not encountered this error before. The only difference with FBEAUT=T should be the additional execution of the subroutine HEXTEND. Since this is a more recently developed subroutine that works specifically on helical objects it is possible that it might have a strange bug that does not always show up. I just ran one round of refinement using the TMV test data set included with the Frealign distribution. I padded the input reference volume and the input stack to 664 x 664 pixels using Spider to reproduce your error. However, the output volume was perfectly fine (I also used FBEAUT=T). Please check all your parameters again. When you change from 2 x binning to no binning make sure you change the scanner pixel size (DSTEP) and output pixel size (PSIZE) to account for the changed pixel size (both values should be halved as you go from 2 x binned to unbinned). The shifts for the segments listed in the parameter file will also be wrong as they are in pixels. You might get away with just another refinement cycle to correct the shifts but you could also generate a new parameter file in which you double the shift values for the unbinned data.

gongpu zhao Wed, 2011-01-12 23:39

Thanks for the advice. I check the parameters and nothing is wrong. the reconstruction without beautify looks right. I find out in the hextend, some values in LSUM(N)=0, which cause B3DV(ID)/LSUM(N) give NaN. This is the reason the final reconstruction can not be visulized. After I fix this i also found out along z axis, some density are zero at certain z value. I am wondering under what condition, LSUM(N)=0. The helical parameter I use is 1 star, alfa=328.87, rise=7.15.

niko Fri, 2011-01-14 11:20

I reran my test script with your values for the helical symmetry but still cannot reproduce your error. Are you sure you are running the latest version of Frealign (check our web page)? If yes, would you be able to email me your script (com file)?

gongpu zhao Tue, 2011-01-18 17:48

Thanks for the quick response. I have tested the updated frealign with my data. The problem has been solved. The reconstruction looks really nice. Thank you so much for your kind help.