Examples of running Phaser in CCP4i
May 2006
These examples are based on the keyworded tutorials from the Phaser website.
There are now similar tutorials available at
http://www-structmed.cimr.cam.ac.uk/phaser/faq.html. The examples were run using Phaser 1.3.2 from CCP4
6.0.
The Phaser documentation is also helpful:
http://www.ccp4.ac.uk/dist/html/phaser.html
Example 1: TOXD with a simple ensemble
- Start the Phaser interface within CCP4i
- Enter the title Example 1 toxd fast rotation search
- Select Mode for molecular replacement to be
rotation search using fast rotation
function
- Select the input MTZ file to be toxd.mtz - the labels
should automatically default to FTOXD3 and
SIGFTOXD3
- Select the spacegroup to be that of the MTZ i.e. P212121
- Set up the details for component #1 in the
Composition of the asymmetric unit folder:
- This should already be set to protein
- The molecular weight is 7139
- The number in the ASU is 1
This gives Phaser an idea of what it is looking for.
- Specify the search model to use for the rotation function in the folder
Define ensembles. An ensemble is Phaser's name for a search
model, which can be made up of multiple coordinate files - to start with with
we will use just one model in our ensemble:
- For Ensemble #1, set the name to
toxd
- Select PDB #1 to be
1D0D_B.pdb NB this can be found in the
CCP4 example data directory $CEXAM/data
- Set the sequence identity to 0.364
This information gives Phaser an idea of how similar the search model
is to the target - alternatively, the expected RMS coordinate error can be
used instead.
- In the Search details folder tell Phaser to Perform
search using toxd i.e. the search model
that you just defined.
- Click on Run
What to look for in the output
The Z-score for each solution is used to get an idea of how likely that
solution is to be the true solution. The section
How to
know whether Phaser has solved it in the Phaser documentation gives an
indication of appropriate Z-score values for rotation and translation
functions.
(The Z-score is the number of standard deviations that the solution
peak height is above the average peak height.)
This job should create three output files:
- .sol conatins the top solutions from this step
- .rlist contains the top rotation solutions, in a format
suitable for input to a translation function search
- .sum contains a short summary of the task run and the
key results from each of the Phaser steps.
Example 2: extending the ensemble
We can improve our search ensemble from example 1 by extending it to
include more information in the form of additional coordinate files.
Phaser will take all the models within the ensemble and merge them into
an averaged model. The idea is that parts of the model which are similar
will be emphasised while dissimilar parts will be weighted down.
In order for Phaser to be able to combine models within an ensemble it
is necessary that the PDB files are properly superimposed. Within CCP4i
you can use the Superpose Molecules task (in the Coordinate
Utilities module) to do this. It is recommended that you also check
the superposition using a graphics program.
- In the Phaser task window, set the title to Example 2
toxd fast rotation search with extended ensemble
- In the Define ensembles folder add a second PDB file
- Click Add superimposed PDB file to the
ensemble
- In the new line labelled PDB #2 that
appears, select the file 1BIK_2_1D0D_B.pdb
- Set the sequence identity to 0.377
- Run the task
Note that if the coordinates are not superimposed (for example if you
use the file 1BIK.pdb rather than the superimposed coordinates
1BIK_2_1D0D_B.pdb) then Phaser will probably stop with a fatal error.
At the end of this run you should get a single rotation search solution
with a better Z-score than any of the solutions from the previous
example using just a single model in the ensemble.
Example 3: TOXD translation search
Having obtained a rotation function solution in example 2, we can
now use this to perform a translation search to get a complete
solution.
- In the Phaser task window, set the title to Example 3
toxd fast translation search with extended ensemble
- At the top of the window, set the Mode for molecular
replacement to translation search
using fast translation
function
- In the Search details folder (below the ensemble definitions)
select the .rlist file from the previous run - this means that the
solutions from that run will be used in the translation search.
- Run the task
Looking at the .sol file indicates a translation function Z-score of 9.4
- since this is higher than the value of 8 recommended by the documentation,
it is likely that this is the correct solution.
To see the difference that using the two model ensemble makes, try
running the translation search with the simple ensemble from example 1. There
are more solutions however Phaser still picks out a solution with a Z-score
of 9.4.
Example 4: Beta-Blip using auto search
This last example is a more complex example which uses Phaser in its
automated search mode. This mode performs rotation and translation
function steps followed by a packing step to check for clashing solutions.
BETA/BLIP is the beta-lactamase/beta-lactimase-inhibitor complex. The
BETA component constitutes 62% of the scattering material and is relatively
easy to locate, whereas the BLIP component is much harder to locate. However
once Phaser has located the solution for BETA, it is able to use this
information to find BLIP.
- Start the Phaser interface
- Enter the title Example 4 Find BETA and BLIP using auto
search
- Select Mode for molecular replacement to be
automated search (this should be the default)
- Select the input MTZ file to be beta-blip.mtz - the
labels should automatically default to Fobs and
Sigma
- In the Define data folder, select Run Phaser
with the mtz space group and enantiomorph (if
applicable). This is because although the spacegroup recorded on
the mtz file is P3221, in this case the other hand is also a
possibility.
- Set up the details for the BETA and BLIP components in the
Composition of the asymmetric unit folder:
- For Component #1, set the values to:
protein, molecular
weight 28853 the number in the ASU
is 1. This is the "beta" component.
- Click on Define another component and for
Component #2 set the values to:
protein, molecular
weight 17522 the number in the ASU
is 1. This is the "blip" component.
- Specify the ensembles in the folder Define ensembles. There will be
one ensemble for each of the components.
- For Ensemble #1, set the name to
beta
- Select PDB #1 to be
beta.pdb with sequence identity
100
- Click on Add ensemble
- For Ensemble #2, set the name to
blip
- Select PDB #1 to be
blip.pdb with sequence identity
100
- In the Search details folder, set up the searches to be
performed.
- Choose beta with the number
of copies set to 1
- Click Add another search - choose
blip with the number of copies set
to 1
- Optionally: in the Additional parameters folder,
check the box for Permute search set, and
select on. This means that both search orders
(BETA first, BLIP second and BLIP first, BETA second) are tried.
(In practice this option is not necessary for this case, as the
larger molecule should be easier to find first.)
- Click on Run
Note that it can take some time to run this example, and that the
resulting log file can be very long.
At the end there are a number of output files - as well as .sol, .rlist,
and .pdb files there should also be:
- mtz which contains "sigmaA" type weighted Fourier map
coefficients for producing electron density maps for rebuilding.
Beyond the examples
Phaser supports a large number of difference options and protocols.
Some options that might be of interest include:
Different ways to specify composition of the asymmetric unit
In the above examples the composition was specified in terms of the
molecular weight of each of the components. However it is also possible
to use:
- sequence (in fasta format)
- percentage scattering
Different ways to specify models in ensembles
Coordinate files are common ways to specify molecular replacement
models. However Phaser allows specification of models in other ways:
- low resolution electron density map - in this case the density must be
cut out and converted to structure factors in a large cell. It is
necessary to specify both the extent of the cut-out region of density
and the centre of this region. For a map, the RMS value does not have
the same physical meaning as when using atomic coordinates, instead
it is used as an indicator of how the accuracy of the calculated
structure factors drops off with resolution (see the Phaser
documentation for more details)