It may be due to change in crystal packing. Did space group also change? Hkl can process the data, check Mathew's coefficient by using merged or unmerged mtz and Molecular weight of ur protein. Its not unusual to see space group change upon soaking of crystal with ligands..

Shivendra Pratap :

Thanks for your answer. HKL results showed that the space group didn't change. But the completeness of the several datdaset is about 95%-98%.

I used CCP4 to convert .sca to .mtz, and then check the Mathew's coefficient. The results showed only 2 subunits. And the model I usually used cannot fit the electron density map even though i changed the homotrimer model to dimer. The R factor after Refinement is about 0.5...Usually it is about 0.2...

If space group is tetragonal or hexagonal then check for possible twinning using L test. Also do molecular replacement with cordinates of your monomer from trimer. Dont use dimer as serach template. If still face high b factor then use buccaneer or auto build just after molecular replacement. Use phaser instead of molrep with search in all possible point groups.

Dear Dr. Shivendra Pratap ,

Thank you very much for your advise. The space group is cubic ( P213), I checked the L-test, L statistic = 0.491 (untwinned 0.5, perfect twin 0.375)

Yes, the monomer can fit the map well. So in this case, is it possible to get the homotrimer? I want to check if all the three chains holding the product-analogue.

I'm also confused why this time the dataset cannot fit the homotrimer template as I used before.