P212121

Phil Evans is the programmer behind Scala which is used for scaling and merging. The presentation and slides were given at the Daresbury Laboratory, UK in 2001. This presentation gives a behind the scenes look at the various inputs and outputs of Scala. I would classify this as a though introduction to scaling and merging.

The presentation runs 52 minutes in length. Another aspect which is refreshing is that Phil simply states what he knows and does not. I for one have been to too many presentations that sidestep perplexing issues and questions.

Questions addressed:
Why are intensities on different scales?
What is the real resolution of your diffraction?
Why you should collect differently for phasing and scaling?

Notes of interest:
Turn B-factor on during Scala if you believe there is a large amount of radiation damage
Rmerge and multiplicity are inversely related
-the two references mentioned relating to this point are Weiss and Diederiches

My only compliant would be that the questions are difficult to hear at times.

Many people that are new to crystallography have a hard time ‘visualizing’ what is occurring between real and reciprocal space. The most intuitive and user friendly program that I have found to help with the problem is XRayView 3.0, which is free to educational institutions.

What concepts does XRayView touch upon?

The software uses an interactive gui to introduce the concepts of X-ray diffraction by crystals, including the reciprocal lattice, the Ewald sphere construction, the wavelength dependence of the reciprocal lattice, primitive and centered lattices, systematic extinctions, rotation photography, space group determination and the alignment of crystals by examination of reciprocal space. Laue cones, photography and group symmetry are also covered with this software.

XRayView is available for IRIX (is anyone still using SGI?), Windows, Linux and Macintosh.

Information about using XRayView can be found here.

The program comes with a number of exercises which are worth working through. In addition, I would recommend simply ‘playing around’ with different parameters, noting the effects of each adjustment. The citation for XRayView can be found here.

The process of indexing is determining the space group from your diffraction pattern. The points of intensity that are hopefully present in your diffraction pattern, I will refer to as spots.

The following are the very basic steps needed to index your diffraction pattern. You made need more advanced commands depending on your data in which you should refer to the ipmosflm documentation.

The process of indexing with ipmosflm loosely moves from top to bottom of the gui.

Initially, you should cover the beam stop or any other shadows on your diffraction image. Use Beam / mask images, which is the bottom button of the gui.

1) Find where the spots are located
-click Read Image (can select an image), Find Spots
2) Move to another image to find spots
-click Read Image (select a different image), Find Spots
(you may get better results by selecting images located away from each other)
(if you have between 200-700 spots then proceed to the next step)
3) Select images to index
-click Select images and select the images from the corresponding number listed
(the program usually detects which images it should use, but best to make sure)
4) Autoindex
-select the best solution, lets start with the default
(you need to type both the solution number and then enter the space group)
(I have found it best to select lower symmetry since it can be increased later)
5) Predict the location of where the spots should be located based on your autoindexing
-click Predict
(if they do not match then you could have a number of issues: 1) wrong space group 2) cracked crystal 3) wrong beam center 4) wrong rotational parameters)
-click Clear prediction (remove the spots that you just predicted)
6) Estimate mosiacity
-the lower this number the better (looking for number around .4-1.2)
7) Refine Cell
Integration
(keep an eye on the integration despite it maybe boring sometimes the crystal may drift during collection or some other issue)

Rarely will you ever have to significantly vary from the default inputs. If you are new to using ipmosflm, I would recommend simply working with the default options to gain a sense of how the program functions. Followed by examining what each option means – ie. don’t become intimidated by all the options to the point of not giving it a try (as long as you have images you can always start over).

The world of graduate school is such a unique subculture that it maybe best described by comics. If you (undergraduate, graduate or professor) are involved in research at any academic institution, you should find something that you will enjoy.

I have included a link to the most popular 200 comics, here.

***see comments for now even a faster way***
Electron Density Server (EDS) provides the scientific community with a service for evaluating the electron density (and, indirectly, some aspects of the model quality) of crystal structures deposited in the Protein Data Bank. The reference for using the server can be found here.

Why would being able to quickly view electron density maps be helpful?

Allows others to gain a deeper understanding/appreciation of your interpretation of how the structure should be built. I have read papers which state something along the lines of “density was missing so I can’t build a given loop.”

Well, how much density was missing? Do you know the carbon alpha locations, but not the slide chains? etc…

Overall, this tool allows for a method to validate a give crystallographic structure. A more complete discussion of the need for the deposition of structure factors can be read in the above reference.

Here are the steps to use the server (it is not hard):
1) submit desired PDB ID into server
2) Left hand side is a ‘Downloads’ section
3) Click on ‘Download maps’
4) Select map format
-if you are using Coot then download either the ccp4 or cns format
5) Select type and download
6) Double click on .gz file, which will cause the files to open in a folder
7) Highlight desired file then click ‘extract’ usually in the upper part of the window
Finally, Open Coot -> File -> Open Map…

If you need the coordinates (where the atoms are located, sometimes called a pdb file since that is its extension) you can download it at the EDS.

There a number of neat options in the EDS, but they will have to wait for another post.

Stephen Curry summed up the real life application of the EDS quite well:
Thanks to the Electron Density Server, it took me all of 30 seconds to go from “Hmm, I’d like to check that out” to “Oh, I see what they mean.” On the EDS web-page you simply enter the PDB identifier for the structure (taken from the paper) and it immediately serves up a package of files that, once unzipped, lets you fire up the molecular graphics program O. You can then get straight to work: in the O session the structure coordinates and maps are already loaded. The EDS is a fantastic piece of work.

Kind of sounds like a clip from an infomercial.

As a side note you can check out Stephen’s blog here, which is an interesting blend of real and reciprocal space.

Mosflm was originally developed by Andrew G.W. Leslie and is now being updated by Harry Powell at MRC Laboratory of Molecular Biology.

Mosflm can process diffraction images from a wide range of detectors and produces, as output, an MTZ file of reflection indices with their intensities and standard deviations (and other parameters). This MTZ file is passed onto other programs of the CCP4 program suite (SORTMTZ, SCALA, TRUNCATE) for further data reduction (reference).

In addition, Harry has a page dedicated toward help and guidance using mosflm which can be found here.

To get started using mosflm (assuming mosflm has already been installed) you have to open a terminal then type ‘ipmosflm’ or ‘imosflm’ (you don’t need a comm file if you are using iMosflm) depending on which version you would like to use. Here is a list of the basic linux commands that you will need to use the program.

This information is also available in the MOSFLM 7.0.4 User Guide. My goal here is to have a comm file that you can simply copy/paste/edit. A “comm file” is a file that allows mosflm to understand inputs specific to your experiment that may vary between users (such as what you named your files or where they are located).

A comm file is simply a text file, which can be generated by typing ‘edit’ into your terminal. Once completed save this file in the directory in which you plan to run mosflm. Save the file with the name ‘comm’ so it is easy to remember when you need to refer to the document using mosflm.

If you would like more information about creating a comm file, you may find this link help.

Below is a hypothical example of what your comm file should contain:
detector marccd (type of dectector used during collection)
findspots threshold 15 (threshold to search for spots this is a good default, but can adjust later)
synchrotron polarisation 1.00 (wavelength used during collection)
directory /home/data/lyso (location of the data)
template lyso_### (what the data files are named)
image 1 phi 0 0.3 (the number of the 1st image, usually 1 and the step size along phi in this case)

The # symbols stand for any number which allows all your images to be read into mosflm.

This is a blank comm file that you will need to fill with your experimental data:
detector
findspots threshold
synchrotron polarisation
directory
template
image 1 phi

You may also need other inputs depending on if your image header was correctly formatted which would include:
distance 120.00
beam 84.444 84.824
pixel 0.0792

How do I prevent condensation on cover slips?

Condensation on your cover slips can easily ruin your crystallization setup. The fact that this may occur after a number days/weeks/months of protein preparation can make this event especially aggravating.

I recently came across a number of posts by Partrick Stewart from Douglas Instruments. I have combined and modified the posts and hopefully provided a sensible thermodynamic explanation. Finally, I have also included a tip to prevent condensation from occurring.

1. If you have condensation, then you MUST have a heat flow, where the heat is flowing to where the condensation occurs.
2. The solutes in the reservoir will stop condensation due to minor heat flows. The more salt etc, the less condensation on the tape.

I find the easy way to think about condensation is to remember that wherever there is a heat flow, that flow carries moisture with it. (Or you could think of it as the moisture carries the heat, although obviously it is the heat flow that drives the process.)

Therefore, when you put a warm plate in a cold-room, heat will flow UPWARDS from the reservoirs into the air above the plate, and you’ll get condensation on the tape or cover slip.

The simple solution is to put a (warm) book from your office on top of the plate when you put it into the cold room. Now you have a heat flow DOWNWARDS from the book, through the plate and into the cold bench.

This method is guaranteed to prevent condensation – in fact it will remove condensation if you have it – just put a warm book on the plate.

I haven’t experimented with the size of book – it could be that I have been actively drying out the drops, which may not be so good!

And I’m sure that other objects would work well although I have only tried books. They seem ideal because they have a reasonable thermal mass but quite low thermal conductivity.

Finally, you will have a use for those Harry Potter books.

Macromolecular crystallography is a biophysical technique used to understand biological molecules such as proteins, viruses, RNA and DNA to near atomic resolution. This high resolution helps scientist understand the mechanism by which these macromolecules carry out their functions in living cells and organisms.

Macromolecules can be generated from recombinant technology, synthetically developed or from a natural source, all of which require purification. Macromolecules can crystallize under the right conditions to form repeating units in a regular 3 dimensional lattice. Once crystals have been achieved, X rays can then be diffracted by the atoms in equivalent positions in the crystal lattice that results in sharp intense spots (called a diffraction pattern). The macromolecular structure can be determined by the analysis of the intensities and positions of the diffraction spots.

PDBalert is a free web-based automatic system that alerts users as soon as a PDB structure with homology to a protein of interest becomes available. Users need to simply upload their personal protein sequences of interest. Once a week, when new proteins are released to the PDB database (on Wednesdays) even if they are on hold, PDBalert compares the new structures with the users’ sequence(s). When a significant match is found, the user is alerted by an email containing a link to the search results. Reminder: have your protein sequence(s) of interest ready when registering.

Why is this application helpful to a crystallographer?

This web application will automatically email you if a protein related to one of your current projects has been deposited in the PDB. PDBalert removes the need to regularly check the new PDB entries and will also highlight relationships that are more distant than what would appear by using the simple key word search in the PDB. This information could help you with phasing via molecular replacement or let you know if you have been scooped.

How is the search performed?

The search is done using the HHpred, which is a remote homology detection server (reference). A pairwise comparision is performed using a profile Hidden Markov Models (HMMs) (reference).

One drawback maybe that you have to register to use this web application. However, since they need to email you if a PDB is deposited of similar sequence then it makes sense that they need some sort of registration process.

The PDBalert system was developed at the Gene Center of the University of Munich. The most recent publication on PDBalert can be found here in pdf format.

As a side note it is nice to see some programs related to crystallography using Ruby instead of Fortran.

How can I quickly find an annotated sequence of my protein?

Try using Pfam 23.0

Click on ‘VIEW A STRUCTURE’
Enter the PDB code (if you would like to simply test this without a code, you can click on ‘example’)
Click on ‘Sequence mapping’

Click on the link under ‘UniProt ID’
This will take you to the UniProt summary page.
To get all the information about your protein of interest.
You will see the following and click on the id (see below).
This is the summary of UniProt entry ‘CLICK ID”

General or sequence annotation information and scan the references.

Why is this tool great?

1) It will save time by taking advantage of the PDB ID which you probably already have
2) UniPort searches both Swiss-Prot and TrEMBL (so you don’t have too)

You can find out more information about Uniport here.