Several things, first your diagrams do not have the correct ratios of any unit cell dimensions, so I am not sure why there is the distortion. But it is so, and therefore not clear in any case which axis is which.
Second you didnt say how you solved the structures. - if it was by molecular replacement de novo, or by solving for the heavy atom positions you could have picked a different origin in the space group. There are 8 such alternate origin positions and it appears from your diagrams that some are exactly that- a different choice of origin. -You would need to select the same origin. Third if you switch h and k in manner you suggest you would have generated a left-handed data set with respect to the other sets and that could cause problems. You would need to switch h to the new k and k to be the new -h. Hence if your diagrams for the last three cases you said were the issue have c in the same direction (coming out of the page) this would exactly explain the 90 degree problem; namely the switch of h for k and k for h, - and the diagrams that dont reflect the unit cell dimensions -but some different reference frame. the best Bob stroud > On Dec 30, 2014, at 10:52 AM, Kay Diederichs <[email protected]> > wrote: > > Hi, > > I have difficulties to comprehend your description. I understand > - most crystals grow in P21212, with cell a=140 , b=110, c=44 > - some crystals grow in P21212, with cell a=110 , b=140, c=44 > Is that correct? If so, can you refine the structures? To what R-values, at > which resolution? > > The sentence "i wonder if i am doing the correct thing by interchanging the > axis" has one answer: as long as you can refine the model, in each case, to > satisfactory R-values and get satisfactory maps at a decent resolution, you > have done "the correct thing"! > > In that case, as Phil says, you might simply have two different crystal forms > - although it seems unlikely that they share the space group and numerical > values of cell parameters. > > But somehow I suspect also some other source of crystallographic confusion. > Are you certain about the 90° rotations? Could it not be 180°? If the latter, > it may be a case of alternate origins - try CCP4's csymmatch or > phenix.find_alt_orig_sym_mate . > > best, > > Kay > > > On Tue, 30 Dec 2014 16:56:03 +0000, <Thii Chand> <[email protected]> > wrote: > >> Dear Dr.Evans, >> >> Thanks for your reply. >> >> The unit cell parameters are a=140 , b=110, c=44., Hence i am afraid, the >> interchange couldn't have occurred due to similar a and b. I had performed >> dehydration using PEG in the said case. I agree that dehydration or soaking >> with ligand can hardly lead to the interchange of axis.But, what i find >> difficult to understand is how come only in three out of nice >> structures,interchange of cell axis occurs, where i had maintained similar >> experimental conditions for all the nine cases. Is there any logical >> explanation for this or am i doing something wrong? >> >> The sp group are P21212 in all the cases. >> >> Thanks in advance >> >> Regards, >> T.chandran the very best! Bob skype: rstroud9502 mobile phone: 415 987 7535
