BUSTER tutorial 4

Rigid-Body Refinement of a Molecular Replacement model

This tutorial illustrates the use of BUSTER for Maximum Likelihood rigid-body refinement of a model as obtained from a Molecular Replacement solution.

The structure of the SCAFet lectin from Scilla Campanulata bluebell bulbs was solved by Wright et al. in 1999 at 3.3 Å at room temperature [1,2,3]. There are 6 monomers in the a.u., arranged in a doughnut-shaped hexamer. The structure was deposited with PDB accession code 1DLP.

A molecular replacement search was run against a new low-temperature 2.9 Å dataset for the same protein; the search model was the NCS-averaged SCAFet monomer from the 3.3 Å structure, AAs 1-235. This search model is fairly complete but has main-chain and side-chain errors in it. The 6 correct solutions of this MR search are used for the rigid-body refinement of the SCAFet hexamer in this tutorial.

What does BUSTER do here

1. cycles 1-10: BUSTER refines then the positions of the atoms, each of the six monomers being treated as a single rigid body; only the positions of the atoms are refined, individual B factors refinement is turned off. The weight for the non-bonded contacts has been set to zero to prevent clashes due to imperfections of the MR search model. No missing atoms are declared.

Input Preparation

A few parameters need some attention: see the simple quick input guide

Main items of the output

1. R-factors plots:

   The R values at cycle 0 are between 38 and 54%, across all resolutions.

   The R values at cycle 10 show a marginal improvement at medium-high resolution, mainly due to the scaling and a more adequate error model; the low resolution values are worse, due to the fact that the model after rigid-body refinement now clashes with the bulk solvent mask. Upon entering a new cycle of rigid-body refinement the initial masks for the bulk solvent will be more accurate, and the initial R factors improve the ones shown aside for cycle zero.

2. Correlation coefficients plots: the correlation coefficients curves are a good tool to monitor the improvement in the structural model during refinement and MaxEnt completion; they are independent of the overall scale factor but do depend on the values of the relative scale factors between the partial and missing structures.

   The (F_obs,F_frag) and (F_obs,F_calc) curves only differ at low resolution because in F_calc there is a contribution from the bulk solvent, while in F_frag only the atoms in the PDB model for the partial structure are taken into account.

   Most importantly, the (F_calc,F_expct) curve depends on the imperfection parameters that parameterise the BUSTER internal error model. The larger the internal estimate for the error on the calculated F, the more this CC curve departs from unity. A comparison between the (F_calc,F_expct) and the (F_obs,F_calc) correlation coefficients curves can inform as to the adequacy of the BUSTER internal error model: if the latter is correct, after the first cycle the two curves should be close to one another.

   The (F_obs,F_obs+d) curve is a measure of the noise on the data vs. resolution. This correlation coefficient is lower than unity when the noise on the data becomes large (typically at high resolution, where the I/s(I) is lowest).

   • It is useful to compare the Correlation Coefficient curves between cycle 0 and cycle 10:

     Cycle 0: At the outset, you can see that the ``observed correlation'' (Fobs,Fcalc) is slightly lower than the ``predicted correlation'' (Fcalc,Fexpct). The initial error model is overoptimistic (especially at high resolution). This is because the initial imperfection parameter for the partial structure and the solvent are too low, (the starting partial structure is VERY imperfect).

     Cycle 10: At the end of the refinement the medium- and high-res CC values are closer to unity than they were at the beginning: this is due to the improved placement of the partial structure. The imperfection Bfactors for the various model components (solvent, partial structure and missing atoms) adjust so that the ``predicted correlation'' (Fcalc,Fexpct) keeps in good agreement with the ``observed'' one, (Fobs,Fcalc). Still the lowest-resolution CCs are not optimal. They will improve upon entering a new cycle of rigid-body refinement: the initial masks for both solvent and missing atoms will be more accurate due to the improvement brought about to the partial structure by the rigid-body refinement.

Final Results

• The coordinate files of the refined partial structure are in the PDB file: shell.01/SCAFet.0.final.pdb.

  Inspection of this PDB file together with the starting strcutrue and the fully refined SCAFet hexamer confirms that the shifts were in the correct direction.

• The structure factor amplitudes and phases are in the file shell.01/SCAFet.0.final.mtz.

References