Back to the tutorial main page

REFMAC5 Tutorial - Part 2

TLS Refinement

In this tutorial, we will use REFMAC5 to refine the model of a molecular chaperone with PDB code 1n5b, treating each chain as a separate TLS group. The asymmetric unit contains two dimers, chains A + B and chains C + D. Data is in P212121, collected to 2.0Å.

A basic working knowledge of the CCP4i2 GUI is assumed.

Before We Start

Required files:

Please click these links to access the files and save them to your computer.

Contents


Simple Refinement

Start by loading the model and data into CCP4i2, and running simple refinement:

Automatic TLS Refinement

Now re-run the job, this time enabling automatic TLS refinement:

Automatic TLS Refinement with Custom Weight

In this case, the model geometry is not very stable during refinement, as can be seen in the graph of rmsBonds per cycle. This is due to the auto-weighting changing the geometry weight during the refinement process. Consequently, we shall now re-run the job using a fixed weight.

TLS Refinement using pre-defined TLS groups from TLSMD

So far we have only used TLS groups as automatically specified by REFMAC5. It is possible to specify custom groups for TLS parameter refinement. It is possible to manually specify TLS groups (see below), as well as TLS parameter estimates. One automated solution is to use the TLSMD server (http://skuld.bmsc.washington.edu/~tlsmd/).
The file 1n5b_tlsmd.tls was created using the TLSMD server.
We shall now re-run the job using these custom TLS coefficients, still using the weight 0.1.

Compare Results

Consider the results from the four different refinement runs.
How do they compare?
Why do the runs involving TLS refinement start with such high R-factors?
What are the pro's and con's of each refinement protocol?
Which refinement protocol is the best, in this case?



Extra Information


Back to the tutorial main page