First, you should examine your data for lower symmetry group and twinning. What program which used by data collection and preprocessing?

This data were collected at DESY and automatically processed with XDS. We reprocessed all the images with the XDS package and ended with the same solution. Using iMosflm gave the same result. Twinning is absent and the software so far did not allowed us to use a different symmetry group.

If it can be useful:

- By eye the resolution looks overestimated, but the statistics are very good down to 2.3 Angstrom.

- The unit cell is huge (65.650 142.300 316.350)

Which SG would you suggest?

Hi Nakia,

Few questions before I will move to the suggestions.

How many chains you got in C2 space group?

Does your protein have oligomeric states?

If it forms dimers in C2 space group, are the dimers formed by crystallographic 2-fold?

If your protein tents to form dimers, than probably one of your crystallographic axis is 2 and not 21.

When you run Phaser you can search solution in all space groups from the same point group, in your case 222. The program will try to solve the structure in 8 different space groups. If you will not solve the structure in 222 point symmetry, then most likely you have a pseudo 222 symmetry. Your real symmetry is P2 or P21 with the strong 2-fold non-crystallographic symmetry (NCS).

To find the real 2-fold axis, you will have to process your data in P1.

Because, all angles are close to 90.0 degrees, you will have to merge your data in three different settings: abc, bca, cab. The one which will have the lowest R merge is the true P21 space group. Then you can use your model to solve the structure in this P21 set.

If your protein has multiple domains, probably, the conformation of your protein in C2 is very close to the search model you are using. It is possible that in your high resolution data (?P212121?) the conformation is different. You may need to split your model and run molecular replacement for each domain.

Even if you solve your structure in P212121, you may need to change the symmetry to lower P21 space group if your R factors will not drop during the refinement.

Sometimes things are more complicated than they look from the first sight.

Good luck!!!

Hi George Minasov,

thank you for your answer. We solved the structure with the higher resolution dataset last week. What you are saying is actually correct, we solved with P 2 21 21. This has been quite puzzling, because we tried to find solutions in the same point group at the very begin, with no results.

We are currently trying to understand why it did not work the first time we tried, and why it worked now.

You did a nice deduction by the few information I provided by the way, congratulation!

Nakia,

If you would like to look professional in crystallography, I would recommend to reindex your data in the proper space group, P21212.

there are only P222, P21212, P2221, P212121.

In the case P21212 and P2221 you can reindex your data to have properly chosen 21 and 2 fold axes.

Good luck!