Map interpretation, a practical example: P2 myelin experimental map.
Introduction: In this practical, we will travel through an experimental electron density map. The structure was solved at a resolution of ~3 A, using essentially one derivative and anomalous scattering. You will, therefore, see what a low resolution, moderately well phased map looks like, and especially the sorts of errors that occur.
The practical takes the form of a guided tour, where you click on a macro on the 'user menu' to go to the next step. If you make a mistake and want to go back a single or few steps, just activate the relevant macro again by typing it in the terminal window.

Getting started: Log onto your workstation, start the Terminal application, set the directory to where you plan to work (e.g., lets say you are on box #3) and create a directory where you will do the exercise

% mkdir 3
% cd 3
% mkdir p2map
% cd p2map

Copy over all the files you need for the tutorial. For students in our X-ray course, get the files like this:
% cp /home/alwyn/p2map/* .
For others, get the compressed tar file from the O distribution site.
Start O using whatever method you have been taught.
Here is an example of a terminal listing from a Macintosh, running OS X, so yours may look a bit different (on OS X 10.2.4, there is a bug so that there will be a pause on starting the program; CLICK IN THE TERMINAL WINDOW). It has been edited a bit too.:

% ono
Mono enabled only
Gamepad disabled
O > Use of this program implies acceptance of conditions
O > described in Appendix 1 of the O manual
O > O version 9.0.0 , Build 02122x
O > Define an O file (terminate with blank):
O > Menu names are not defined.
O > Enter file name [/Users/alwyn/o/data/menu.odb]:O > Startup file was never loaded
O > Enter file name [/Users/alwyn/o/data/startup.odb]:
O > startup.odb file for O
O > Stereochemistry file was never loaded
O > Enter file name [/Users/alwyn/o/data/stereo_chem.odb]:
O > No codes checked
O > Connectivity used is : all
O > There were 44 residues.
O > 268 atoms.
O > Making visibility data structures.
O > Making visibility data structures.
Fm> Number of mouse buttons 1

To start things off, just activate the macro p2_start like this
O> @p2_start

O will have generated a black window that contains some text. This window is a 3D view of space where you will be able to view and interact with a molecule, or electron density, for example. You cannot see very much in a static view into this window, so O allows you to change the view in a number of different ways. Some workstations may be equipped with dials, usually with 8 knobs, others will not. The dials can be mapped to different functions, and in the bottom left-hand corner, coloured in magenta, there is a description of what the dials on the dial box do (if you have dials attached). The dials, for example, allow you to rotate/translate, zoom and slab the view of the objects you are looking at.

If you have a 3-button mouse, Rot X and Rot Y can be simulated using the right-hand mouse button. Pressing the button down and moving the pointer horizontally simulates Rot Y, and moving vertically simulates Rot X. Pressing the right-hand button and middle button simultaneously and moving the pointer up and down simulates Rot Z. The middle button alone controls the Slab and Zoom. Pressing the middle button and moving the pointer to the left decreases the thickness of the slab you are viewing, and to the right increases the slab. Moving the mouse up and down with the middle button pressed simulates the Zoom. The left button is used to ‘pick’, either identifying atoms, or activating menus.

The Apple Macintosh comes with a single-button mouse. The mouse is very elegant so that makes up for it to some extent. The single button is used to pick and control the Rot X and Y axes. To zoom and slab, hold down the <alt> key and the mouse button together.
The arrow keys on the keyboard can also be used as input. The vertical arrow keys change which input is to get mapped, and the horizontal pair of arrows will generate a value from the corresponding input device. The text ‘->‘ in the lower left dial allocation window indicates the currently active input. This is not a very intuitive way of changing the view for example, but is useful with some commands. You will also see a window called ‘Dial Menu’ on the left of the screen. If you click on one of the panels in this window, that dial will be active for arrow key input.

At the top of the screen are some lines of text, the function of which will become apparent later. In addition there is a set of pull-down menus (in magenta). To see what is in a menu, click on one of the menu titles with the left mouse button. A list will appear below the title. Sometimes an item in this list will generate another window or list. In the menu 'Menus' (far right), each of the items in the list will generate another menu, for example, but the first three in this pull-0down are already opened for you.

By typing @p2_start you have activated a set of instructions contained in the file of this name. You will see a window called the 'Object menu' that has been opened, another one called 'User menu', and a thors called ‘Dial Menu’ that was mentioned earlier. The 'Object menu' will get bigger and smaller as we go through the exercise, but it always contains a list of objects that contain different sorts of 3D graphical data. Clicking on one of the names will either make it appear or disappear. At the beginning, there are only 4 objects in this list. One shows the unit cell, the other is a so-called skeletonized electron density. The 'User menu' contains a set of macros that you will use during the exercise. You just have to click on one to activate it. Their function will be explained later on, but for now notice the last one @p2_1 This is the one to click on when you are ready to advance.

You are likely to stop the program sometime during the tutorial. Your status will be automatically saved for you as you work along. This is a file containing a copy of your O database in the Unix file system, and it has been given the name p2.o On subsequent starts, all you need to specify at the first prompt is the name of this file, and a series of <return>'s

During the practical you will be working with electron density maps that will be read in for you when you start the program.


How many molecules: After activating the @p2_start macro, you will see a representation of the electron density that we call a skeleton. The skeleton represents aspects of the connectivity in the electron density map as well as a certain level. It has been generated from the electron density map and describes a set of atoms with a set of connections. The skeleton atoms have been coded depending on how well they are connected. This coding represents a first division into possible main-chain and side-chain atoms. In this picture, we are looking at main-chain atoms (in object MC).

This molecule was crystallized in spacegroup P212121 with cell constants a=91.8 Å, b=99.5 Å, c=56.5 Å. The following picture is taken from International Tables to show you the symmetry axes (all are screw axes) for this spacegroup.

The red box drawn by O represents the unit cell, and the yellow lines are the various screw axes (notice that they do not intersect). Change the zooming so that you can see the complete cell. Try counting how many blobs you can see within the cell. Spacegroup P212121 has 4 repeating units in the cell so how many molecules are there in the asymmetric unit? When you think you know the answer, put something on top of your computer screen.

@p2_1 will take you to a copy of one molecule that I call A. Notice the contacts it has with other molecules in the cell. @p2_2 takes you to molecule B, which is quite close to A. There is roughly the same number of skeleton atoms associated with this molecule as with A. Now @p2_3 takes you to the third molecule. How does the skeleton look to you, is it as 'dense'? What does C interact with? Put something on the top of the screen again when you have an answer.

The next macro @p2_3a generates the position of symmetry related molecules. A new molecule has been loaded into O that consists of just a few atoms, one roughly at the centre of each molecule. Two more objects have now been created, one of the new atoms, and the other of all symmetry related atoms that are within 100 A. In the Object_menu window, set the object called 'sym' to off Identify the 3 atoms in the object BIT (you may have to click MC off) one at a time and then turn SYM on again. Every blob should now be labeled.


Connections to A: The next 2 macros (@p2_4 and @p2_5) will show you some of the connections to molecule A. This is how crystals get made, and without these interactions crystals could not be formed in 3D. When trying to trace a structure in a map, it is important that you do not wander off into another molecule.


Secondary structures: As you probably know, proteins are built up from secondary structure elements called a-helices and b-strands. Often the b-strands abut each other creating a sheet, held together by hydrogen bonds formed between main-chain atoms of each strand. Activate macro @p2_6 and try to find a b-sheet. If it is difficult, use the macro @show_beta It will generate an object that looks like a series of tubes. Each tube represents a place where part of a b-strand can be fitted to the electron density map (Kleywegt & Jones (1997), Acta Cryst. D53, 179-185). Delete the object BETA when you are done (pull-down Display/Delete_object).
Macro @p2_7 will show you a lovely sheet, made up from 6 strands. Notice how the sheet curves? This is very common in such sheets. Any idea why?

There is another sheet somewhere; can you see it? Hit @p2_8 when you are ready and start looking for an a-helix. There isn't so much helical structure in this molecule, so activate @show_alpha if you have problems. How many helices can you see? Activate @p2_9 to see where the easiest one is located. Notice the little balls of density; some are quite close to the b-sheet. Why are these here when we are looking for an a-helix? Delete object ALPHA when you are finished.


Some reverse turns: Now we start working along the skeleton, looking for errors and using the electron density to tell us what the problems are. The first macro, @p2_10, takes us to a suitable starting point, where we have a pair of strands connected in error. The 2 strands should be obvious, and the connection is roughly at right angles to the strand direction.

Activate @p2_11 to activate the electron density. Actually, we already read in 2 different electron density maps; both are experimental. One is the MIR phased map, and the other is the result of five cycles of density averaging. Most of the time we'll be working with the MIR map since it is more representative but if all else fails, check out the averaged map, which is much easier to interpret. The macro opens up 2 new windows, MIR and AV, which control how you interact with the maps. Each window has a number of sliders, where you can set the radius for contouring, the density level (in units of the root-mean-square deviation in the map), and the 3 RGB colours. Move each window to where you want them, point at the right green square in each box and drag it. Some of the workstations may not feel capable of displaying a large number of vectors, so keep the radius low (the Linux boxes in our new teaching lab are excellent systems for map work; they can display many more vectors than you can sensibly manage). Set the maps to colours that you like. Suitable levels will be between 1-2 RMSDs.

By clicking on and off the objects MIR_1 &/or AV_1 you should see that the skeleton closely follows the MIR map. The skeleton now has red atoms too. These correspond to potential side-chain atoms. Sometimes, you will see a little separated, unconnected group of red atoms.
Once you feel you are able to control the density sliders, move on to @p2_12

Click off the map objects for a moment and zoom out to see some text. Travel along the skeleton between the 2 points indicated. Use the @next_id macro to specify the next screen centre, i.e. click on the macro, then on a skeleton atom- the density gets recontoured at the new centre. You should be able to see how the main-chain of the b-strand has a zigzag like motion so that side-chain atoms alternate pointing up and then down.

Identify one of the light blue skeleton atoms, and look at the 3rd line from the top of the 3D window. You will see the name of the skeleton molecule (START), that it is a BONES residue of type SKL with a number for the atom name. You will then get the coordinate of the atom, nothing for the temperature factor (B), and an atom type indicator (Z) that has a Z = 3 value. Red skeleton atoms will have Z=2. These internal values indicate to O how to colour the atoms. You can change the status codes of skeleton atoms with the pulldown Bones/'Set to Class' command and then identifying two connected skeleton atoms. As you go along the skeleton, this can be used as a marker of where you have been. Set a piece of the main-chain skeleton to Class 1, and then activate @p2_13.

The colours associated with each class of skeleton atoms are saved in the O database in a datablock called .bones_colour The values we are using colour class 1 atoms as yellow, class 2 as red and class 3 as cyan.

Eventually you should come to the position where I've indicated that there is an error. Now what do you do? Click off the electron density object for a moment and look at the skeleton to get an idea of where the main-chain has come from and where it should be going. Try to see the zigzag nature of the strand, since this will help you to recognize where the side-chains are pointing. @p2_14 will take you directly to the error. Change the slider 'Level' to get the density at a suitable level to better indicate side/main-chain. Usually the main-chain density will be stronger than that of the side-chains, but not always.

It's a good idea to save the state of the O Database quite frequently. This is done for you every time you activate the @next_id macro, but you can also do it with the Controls/Save command. If you crash the program, when you start it again you'll be back to the last time you did a save! IF you crash the program or need to start again, just type the file name p2.o at the first prompt and then series of <cr>s until you get the 3D window back.

Once you have decided what to do, break the bonds and colour things accordingly. Then start the next macro @p2_15
Note that there is no skeleton side chain off Bones 2302 (the skeleton atom indicated by the text ‘No side chain’). It could also be positioned a little better. Use Grab_atom to place it (Rebuild/Grab/Grab_atom .... pull-down), then Bones pull-down 'Add branch' to add an atom. The Grab_atom command is used with the mouse, just ID an atom and drag. The command stays active, use Controls/Clear_flag when you've placed it. There is no need to position or recolour the new, branched atom. It will become red the next time it gets redrawn.
Activate @p2_16 when you are ready. This will indicate the next piece of structure to work on. You will pass through your second reverse turn before coming to the next problem. Try to sort it out, using the tools you've learned about. When you think you have it right, activate @p2_17 Remember to colour your path so you know where you've been. It may be time to sit back and admire your handiwork, so click on the @mc macro to get a view of what you and the program think is main-chain. This is in an object MC and you will not want the more complicated MSC object, so remove it for a while (click on the name in the Object_menu). Your yellow path should be clearly seen in this new object.

The mc and msc macros are a bit more complicated than you might think. Here is the contents of one of them, mc:

copy save_centre .active_centre
copy .active_centre molecule_centre
bon_set start mc 23. 1 3 ;
bon_dr
copy .active_centre save_centre

Welcome to the world of O macros.


A loose end: Macro @p2_18 takes us to a loose end. In a perfectly phased map, at a suitably high resolution, a main-chain skeleton would have just 2 loose ends; the N- and C-terminal residues. The reality of map interpretation is that maps are not perfect, so there will usually be many loose ends. With some experience, a crystallographer will realize that some loose ends warrant more thinking about than others. Why?
Make sure the MSC object is visible. Now work your way along in the direction indicated, adding branch-points where you think there should be a side-chain. What we are looking for now, is an indication of where we are in the sequence. The branch-points become spacers, useful indicators for how many residues there may be in a particular piece of skeleton. Here are the different amino acids


When you get to the ‘*’, try to guess what sort of side-chain it is. Then activate @p2_18a. It will generate a new object showing a single residue, and it fits perfectly! Most people do not know how the different amino acids look on a 3D display system, but there is a command that lets us view each amino acid looks and also shows the most frequent rotamers that are associated with a residue. Go to Rebuild/Database/'Rotamer to show' and then click on an atom in the new residue object that just appeared. You will then see a green object superimposed on the coloured atoms. Two new dial inputs have become active and these will allow you to scan through the different amino acids and their associated rotamers. This command does not work with skeleton atoms; you need to have built at least an alanine residue before you can use it. The command will give you a sense of 'scale' to see how things match to the density. Hopefully, you will see that one residue type matches very well. It is not a particularly common amino-acid so we have some very useful information as to where we are in the sequence of this molecule. Use Clear_flags or NO to terminate slider_lego.

Now activate @p2_19 and work your way along to the next problem. When you get there activate @p2_20

This is the most difficult problem you have encountered so far. It may help if you go into the overview mode, where you just look at the main-chain (click on @mc, switch off MSC and the map). Try to sort it out, and then click @p2_21

Did you correctly recognize that there was another molecule close by? In the next part of the practical we'll work on the a-helix that just came into view.


Directionality: When going through a map, we are on the lookout for two sorts of information: sequence specific (e.g. 'this is probably a tryptophan'), and directionality (e.g. 'this is in the direction of the C-terminus'). If phasing and resolution are good enough, the problem is straightforward: if we can see peptide oxygen atoms, we can decide directionality. Draw how a polypeptide main-chain looks, and then you should be able to work out why this is true.

At lower resolution, we can get directionality because of the 'Christmas Tree' affect. If we take the trunk of the tree to represent the helix axis, in a Christmas tree, the branches have a tendency to point roughly in the same direction. Some might start off pointing in a particular manner and then change direction. The tree trunk is equivalent to the helix axis, and because the side-chains start off pointing in the direction of the CA-CB bond, they will have a tendency to 'hang', just like the presents on the tree. They hang back towards the N-terminus of the helix, so now we can figure out local directionality.

After activating @p2_21 work your way along the skeleton of the helix. Try to guess the directionality.

Now activate @p2_22. The skeleton has a break, but it should be obvious what to do; fix it, colour it and save. Move along to the * The area is a bit of a mess, you will probably have to grab_atom and Bones/branch to fix it up. Then activate @p2_23 and go on a bit further to the next *. Now make an overview again and admire your handiwork. What have you just created? If you are unsure click the @show_alpha macro again (make sure the object ALPHA is on).

Now you can go through the directionality quest again, and hopefully it will be in agreement with the first helix. What do you do if it isn't?


Strands galore: We are making progress!. We have the directionality for a big piece of the structure, right back to that loose end. Now activate @p2_24 and move along the strand, building up your idea for how the molecule is folded. Remember what you learned about strands, the zigzag, up-down side-chains, this will help you as you come to any breaks in the density. Click on @p2_25 when you come to the inevitable break. Once you have fixed the problem, click on @p2_26 and make your way to the next problem. This is one of the worse places in the whole map. Best of luck!

Activate @p2_27 to see what to do. If you did it wrong, please fix it up, there's more to do further along after the next reverse-turn. That will be the second reverse-turn after the helix, you are clearly tracing out an anti-parallel b-sheet. Click on @p2_28 when you reach the * This is the worst place in the map but you can fix it if you remember what you have learned about b-strands. Play with the level in the density slider window; it might help.

Macro @p2_29 will show you how you should have done things. If you made an error, fix it up and carry on to the next *. Now click on @p2_30 It's time to admire the structure again, and the macro will generate the main-chain skeleton object MC. Have you been colouring the main-chain and save'ing?

From the overview, you should see that there isn't very much left to do. Butttttttttttt try to fix it, without going into the neighbouring molecule. Click @msc to get all the skeleton atoms. Macro @p2_31 will explain the way. Please fix it if you had a different idea of how things look. Carrying on to the next * is quite easy, going around yet another reverse-turn. Click on @p2_32 and carry on to the next * going around the reverse turn and down a nice long strand.

Click @p2_33 to make sure we are in synch and try to fix this bit too. @p2_34 explains what you should have done, and carry on up the strand to where you started this exercise. Remember it? The next macro will create the overview for us.


Complete trace: click on @p2_35 to admire your skeleton trace of how the molecule folds in space. There should be just 2 loose ends, if not, go and investigate. One end will be the N-terminus, the other will be the C-terminus. Since we (think) we know the chain directionality, it should be obvious which is which. Now we'll go to the N-terminal region and try to figure out what is what. Click @p2_36 and we have 4 residues that I want you to think about. Try to guess what they are, and feed your guess into the Slider_guess command (go to the terminal window). Here is the sequence of the protein in the 3-letter code:

1: SER ASN LYS PHE LEU GLY THR TRP LYS LEU
11: VAL SER SER GLU ASN PHE ASP GLU TYR MET
21: LYS ALA LEU GLY VAL GLY LEU ALA THR ARG
31: LYS LEU GLY ASN LEU ALA LYS PRO ARG VAL
41: ILE ILE SER LYS LYS GLY ASP ILE ILE THR
51: ILE ARG THR GLU SER PRO PHE LYS ASN THR
61: GLU ILE SER PHE LYS LEU GLY GLN GLU PHE
71: GLU GLU THR THR ALA ASP ASN ARG LYS THR
81: LYS SER THR VAL THR LEU ALA ARG GLY SER
91: LEU ASN GLN VAL GLN LYS TRP ASN GLY ASN
101: GLU THR THR ILE LYS ARG LYS LEU VAL ASP
111: GLY LYS MET VAL VAL GLU CYH LYS MET LYS
121: ASP VAL VAL CYH THR ARG ILE TYR GLU LYS
131: VAL

Here is an example of the command in action

Slid> slider_guess
Slid> There are 131 residues in molecule.
Slid> Estimated sequence: lprq
Slid> Average= 0.54,rms= 0.13
Slid> LPRQ
Slid> Fit 1 0.800 124 CTRI
Slid> Fit 2 0.800 66 LGQE
Slid> Fit 3 0.800 104 IKRK
Slid> Fit 4 0.775 42 ISKK
Slid> Fit 5 0.750 98 NGNE
Slid> Fit 6 0.725 68 QEFE
Slid> Fit 7 0.725 59 NTEI
Slid> Fit 8 0.725 32 LGNL
Slid> Fit 9 0.700 62 ISFK
Slid> Fit 10 0.700 35 LAKP
Slid> Fit 11 0.700 2 NKFL
Slid> Fit 12 0.700 86 LARG
Slid> Fit 13 0.700 15 NFDE
Slid> Fit 14 0.700 95 QKWN
Slid> Fit 15 0.675 37 KPRV
Slid> Fit 16 0.675 77 NRKT
Slid> Fit 17 0.675 127 IYEK
Slid> Fit 18 0.675 63 SFKL
Slid> Fit 19 0.675 29 TRKL
Slid> Fit 20 0.675 51 IRTE
Slid> Do you want to associate with a residue ([Y],n)? n

Hint, my guess is not appropriate here. You might have found it hard to estimate the size, so let's get a bit of structure on which we can hang the Slider_lego command, click on @p2_37. We will now use 'Secondary Structure Templates-SSTs'. Use Density/2ry_unit/beta_5 to choose a small b-strand of just 5 residues. Use Density/Template/Fit_beta5 and then ID the skeleton atom at the letter c. You get the best fit of this template to the map after a complete rotational search about the central residue of the template. Hit YES (Controls/Yes) to accept it. Now improve it a bit more, activate Density/E.D.fit/RSR_group and ID an atom in the template, again use YES to accept the new structure.
Now you have something to give you an idea of size so use slider_lego to help you decide which amino acids we have in this area. Type slider_lego, just ID an atom in each residue and try the different amino acids, and their rotamers.

Once you are happy, try slider_guess again. Hit @p2_38 when you think you know where you are. Did you try your guess in both directions? At this resolution, you CANNOT be sure that the template is pointing in the correct direction. The correct answer is: a is Gly6, b is Thr7, c is Trp8, d is Lys9. Notice that there is no side-chain density for the lysine residue. Why?

Time to start creating the real molecule


Using templates
Directionality
Lets go to the first a-helix, and fit a 7-residue helix. Click on @p2_38, choose a new template (Density/2ry_unit/Alpha_7_resdiues), make a rough fit (Density/Template/Fit_alph7), improve it (Density/E.D.Fit/group). If the helix is in the opposite direction from what you thought, reverse it (Density/Template/Flip).

Previously we guessed the directionality of this helix, but we can do it more quantitatively. First specify that you want to use the MIR map (pull-down Density/ Density/MIR). Activate Density/Helix_tree and then ID the first and last residues in this helix. Now go to the Density pull-down and you will see a new map called LOCAV (local averaging), rip it off, set an appropriate colour, and the level to about 1.8. The hanging side-chain density should now be more apparent, and if you've built it in the right direction, the helix is fitting snuggly in the density. For fun, reverse the helix (Display/Template/Flip). It should be obvious that this is a much worse fit for the CB atoms.


Making a CA-trace
We could carry on here with the edited skeleton that each person or group has made so far, but each skeleton will be different and any problems you encountered with it will be different. Therefore, we will read in a new skeleton that I made from the average map, where I have chosen an appropriate level for the skeletonization, and made 2 small edits. The latter will be obvious when you view it since I have coloured them as class 1 bones atoms.

To start this session, just type @decorate_1 and have a look where I had to make some changes. I have also updated the ‘User Menu’. Click on the top-left squares fo rthe map-sliders for the MIR and LOCAV maps, we’ll not be using them.
Go to Display/sequence and rip off the window by clicking in the little top left red box. Position it somewhere by clicking in the top right square and dragging. The vertical line represents 100 residues, click somewhere on the line and you’ll see the local sequence in a 5-residue window.

The skeleton consists of just an NCS (non-crystallographic symmetry) asymmetric unit, after 5 rounds of averaging. The phases are much improved compared to the MIR map, but the resolution is the same. In this session, we will use tools that work on well-phased maps, or carefully edited skeletons of more poorly edit maps. Type @decorate_2 to generate a CA trace from this skeleton. This macro just activates a command (Trace_ca) that applies a set of rules to the specified skeleton to generate another skeleton called CAT. The latter represents the Central Atom Trace of our structure. In this example, we should see something that looks a lot like a Ca trace of P2 myelin. Take a look at the results, in the CAT object; I have marked problem places in the CAT with red letters. Because it is a skeleton, it is easy to change things: we could reclassify the skeleton atoms with Bone_redefine, move atoms with Grab_atom, or change the connectivity with Bond_make and Bond_break. It would take just a few minutes of work to make a better CAT by editing the starting skeleton, but we’ll not bother.

We can do a bit better by fitting SSTs to the skeleton, and then using these as well as the skeleton to make the CAT. SSTs can be added to the TRACE molecule interactively, or more automatically. To do the latter requires a calculation for how well each template we plan to use fits at each atom in a skeleton object. Click @decorate_3 to start this process for the MSC object for SSTs that correspond to a 7-resdiue a-helix and 5-residue b-strand. It will run for a few minutes and then we can start building.
Follow this example to build the 3 best fitting a-helices automatically:

spr_sst
Slid> What map? [AV]:
Slid> Skeleton molecule [CAT]: av
Slid> Number of residues 1
Slid> Number of atoms 1236
Slid> Secondary Structure [alpha]: alph7
Slid> Number of residues 7
Slid> Number of atoms 35
Slid> How many shall we build? 3
Slid> Density fitting score 13.5119
Slid> Density fitting score 10.8535
Slid> Density fitting score 10.0806
Slid> All SSts built
Heap> Deleted TRACE_CONNECTIVITY
Mol> Database compressed.
Mol> Created connectivity Db for TRACE
As1> vis_obj cat off

The density fit scores are, roughly speaking, the average density at each point in the SST. The third SST has been fitted to just a single turn of helix, and needs it’s C-terminal to be trimmed by 2 residues (Density/Template/Trim). How many of these helices are in the wrong direction?

Now repeat asking for three more helices. One goes into a symmetry related molecule; the other two are skewed across the ‘back’ b-sheet. Notice how most atoms in the helix actually fit to the density; their goodness of fit indicators are not much worse than the correct helices. Remove the wrong SSTs (Density/Template/Remove_SST)

Now repeat for the five-residue b-strand, adding twelve strands. Some of these will be in the wrong direction; you might even see an example of the ‘carbonyl oxygens in the side-chain density result’. A few SSTs need trimming, and one of them needs interactive help (Rebuild/Grab/Grab_group use the mouse to drag and the <ctrl> key/mouse to rotate). Do what you think is best.

So that we can be at the same place, read in my version of what the TRACE could look like with @decorate_4

I’ve made an object of your TRACE and called it YOURS. You might not be able to identify any atoms in your object; do NOT Centre_ID on them.

Now try Trace_Ca again with @decorate_5. Some of the places where there are 3 main chains coming from a single skeleton atom, have now been ignored, but a few wrongly defined skeleton assignments still cause problems; they have ‘*’’s on them. The problems can be fixed either by modifying the AV skeleton or by editing the CAT skeleton. If you modify the AV skeleton, just try Trace_Ca again with @decorate_5. When you are happy with your CAT click on @decorate_6 to see my CAT.

If this were a real structure determination, I would now look over the CAT to see if anything can be improved. In @decorate_7 we will now turn this back into a TRACE, using the main-chain database option in Sprout_CAT (at higher resolution, we would use different options). The old TRACE will be zapped in this macro. The TRACE as a CA object will, of course, look a lot like the CAT. You will see one large chain, residues 6-130, and a few short ones. The latter can be ignored. Now use the @show macro to see how different sorts of amino-acid side-chains fit the density. Click the macro, then a TRACE atom. Five side-chains are show superimposed on one another, each in a separate object called SCXYZ where XYZ is the amino-acid 3-letter code. On the terminal there is a score for how 13 amino-acid side chains fit the local density. Only the rotamers for each side-chain are tested, pivoting about the CA atom of the TRACE. If you go to the Density pull-down and look at the M5 slider, you will see the contours of a mask into which the real-space fitting takes place. This mask depends on the value of the slider in the map being used (in this case the AV map). This gives the user some control over what density is used in the fitting. As an example, what if 2 side-chains are interacting, if you set the AV slider to a higher value, you could stop the side-chain moving into the density of the other.

Go to residue 98 in the TRACE ‘cen_at trace 98’ and tell me what you think this is. What about residue 107 in the TRACE?
Now click on @decorate_8 to continue. This will carry out a similar analysis at each residue on the TRACE, Decor_guess, and takes a few minutes to run. You will then see a preliminary placement of the sequence of P2 myelin onto the TRACE. The optimal path for decorating the TRACE will be written out, and look like this

Optimal path 1 7 11 16 27 48 58 69 81 91 100 113 122

The TRACE has been redrawn to show the alpha helical segments in red, the strands in green. The green text indicates where O suggests you should place the sequence. The red text indicates where the TARCE and the path differ in length. Usually the TRACE will take a short-cut through the density; look at the last 2 strands, for example. The penultimate strand is suggested to end at 118K and there are 2 residues in the TRACE before the first residue in the last strand. However, this residue is suggested to be 122V, which would require 3 resdiues in the reverse turn instead of the 2 we have in the TRACE. The difference between the TRACE and optimal path is indicated by the ‘*1*’ text in red. If the TRACE is longer than the path, you will see a negative sign in red; look at the gap between 9K and 11V, for example. The TRACE has 2 resdiues, the path has only 1 residue, so O draws * -1* here.

A new window will also appear showing the sequence of P2 myelin. If you move the mouse over the vertical line in this window, you will be able to change the view the sequence in a 5-residue window. Click on the >/< characters to go above 100 and to then return below 100. This so-called sequence-slider window is used with the Décor_fix command to correct errors in the optimal path. Set the sequence position you want in this window, then click on décor_fix, then ID an atom in the TRACE. The TRACE atom MUST be one of the helix/beta coloured atoms. A new path is then calculated. For example, if you think TRACE residue 18 is actually residue 12 in P2 myelin, set up resdieu 12 in the sequence slider window, so that it is green, then click on Décor_fix in the User Menu, then ID residue 18 in the TRACE. The path changes to

Optimal path 1 7 12 16 27 48 58 69 81 91 100 113 122

When you are happy, click on YES and O will build everything according to the path. It will use a skeleton to build the gaps between the secondary structure stretches. Here is my example

Optimal path 2 7 12 16 27 48 58 71 82 91 100 113 122
yes
New> Result saved as :1AA
New> What map? [AV]:
New> Map used is AV
New> Skeleton used is [CAT ]: av

I would do a round of rotamer fitting and regularization without looking at the results, and then go through the structure carefully with the Grab_build command. Click on @decorate_9 to do this for us. There are only 2 places where side-chains having good density do not fit very well, residues 4 and 128. My model is from residue 2 through 129, so the end residues will also need to be fitted.

Ask how to get my path through the sequence (Internet users can send me an e-mail, or activate @cheat but Clear_flag first). Once you are convinced I’m right, click YES. You still have residues 4 and 128 to fix since their side-chains are not in density.


Optimizing the fit: Grab_build is a complicated command, controlled by its own window. The first line in its window stops the command, the second line allows you to move forward or backwards in the molecule. BUT the coordinates that you see for the next new residue depend on the fifth line that controls how the chain ‘grows’. The starting grow mode is ‘As is’, i.e. the coordinates as they are in the molecule. We are centred on residue 4, make sure the map is visible. Press the <ctrl>-key and move the mouse,; the residue will pivot around the CA atom. Click on the ‘>’ in the last line of the Grab_build window to select an ‘Action’ that will do a RSR rotamer for us. Once you have the side-chain closer to the correct density, click on this action text. Regularize if need me (first action text), or use the torsion if you really want to fuss (second action text). Then click on @decorate_10 to fix residue 128. Click on the ‘Quit’ text when you are happy.

The end: When you have finished, activate the macro @p2_end
 
I hope you enjoyed the tutorial