DirAx Examples

The examples 1-4 are those given in the paper. The actual results may be slightly different (notably for example1) due to modifications in program details. Only user input is shown in the Examples. 

              ! note:
              ! these are comment lines,
              ! do not type.
Use the command example 'n' to read example data.
You can run all the examples in one command: dirax < ex_solutions.inp > ex.out
and compare the results with the file ex_solutions.out.

0 straightforward
1 fragmented crystal
2 area detector small protein
3 incommensurate crystal
4 one axis much longer
5 twinned crystal
6 inaccurate data
7 not accurate and large cell
8 twin crystal and ill-distributed data
9 twinned crystal
10 area detector data protein
11 two lattices
12 twin. Two solutions in one run

Example 0

ex00 - This is a straightforward model run for a single lattice. (In fact too easy for DirAx.)
dirax ! start program
example 0 ! read ex00.drx
go ! run with defaults
lo ! accept proposed ACL and show H indices
ro ! show cell and [R] and [D] matrices
write ! write file ex00.out for print-out
end ! finish program
Note: without end DirAx remains active and you may continue with other DirAx commands, e.g. read filename, example nn. If appropriate, use default to reset all parameters before a next go.
This example is also presented in the DirAx Course.

Example 1

ex01 - Data from a fragmented crystal. First run with default parameters and accept proposed solution. Inspect the hkl list and note that more reflections could be made fitting with a wider IndexFit: change value to 6 and select the default solution.
dirax !
example 1 ! data from a fragmented crystal
list ! show input data
go ! run with defaults
lo ! note that some refl's might fit with...
indexfit 6 ! ...a more relaxed fitting criterion
acl ! display solutions with new parameters
auto ! select solution with most H refl's
lch invert ! interchange n <--> H
go ! go again with n-refl's only, for other lattice
end !
The alien reflections do not form one lattice and cannot be identified properly.

Example 2

ex02 - Area detector data for a small protein, note narrow phiB region. Set Dmax to 160. The crystal is known to be tetragonal, so there IS something wrong with the data(conversion)!
dirax !
example 2 ! area detector data for a small protein
dmax ! display current dmax
160 ! dmax set larger than usual
go !
lo ! list H and n refls
write !
end !
You might try the default dmax = 80 and see what happens.

Example 3

ex03 - Data from an incommensurate crystal from dr. J.L. de Boer, Groningen. As usual these are very accurate data (from SET4 procedure on CAD4). We know that problems might be ahead so we set LevelFit to 1/3000 (more severe criterion) to discriminate interfering reflections.
Note that super solutions are found with (almost) all reflections fitting. We know however (aha!) that the correct volume is about 100 so we select the appropriate ACL. In the hkl list you see that not fitting reflections are at discrete distances from main lattice points.
dirax example 3 ! data from an incommensurate crystal
go ! run with defaults
acl auto ! first accept 'best' solution
loh ! list fitting reflections
acl 16 ! now overrule superlattice (*)
loh ! list fitting refl's (x)
lon ! list satellite refl's
end !
(*) Note that for this solution alpha = gamma = 90 and a = c, so it is worth trying it. (Moreover the cell volume was expected to be about 100)
(x) Note the extremely small errors (1/err): SET4 data from dr Jan L. de Boer, Groningen, NL.

Example 4

ex04 - One axis much longer than the others (more than 80 A), a classic indexing problem. Set Dmax to 120 (200 will work too, not critical if large enough).
dirax ! one axis much longer than the others,..
example 4 ! ...a classic indexing problem.
dmax 120 ! increase dmax
go ! for the rest use defaults
lo ! inspect list
end !

Example 5

ex05 - Data from a twinned crystal. Default parameters. A super solution is found for all reflections, which is common with real twins.
Note: It it not possible to give general rules for this sort of problems. The super cell COULD be correct (and IS geometrically!) but you have to consider crystallographical aspects. Here we select ACL 18 because this looks promising. Write results to file ex05.out1.
Continue with 'n' (not fitting) reflections only. Now the other lattice is found. Write to file ex05.out2 and compare with .out1 later.
Normally with so few 'n' reflections a sub-lattice is found rather then a congruent lattice. Then you have to search further selectively.
dirax !
example 5 ! data from a twinned crystal
go ! run with defaults
acl 18 ! overrule super lattice solution ACL 25
go ! go again with H-refl's only, for 1st lattice
cell ! cell etc. for 1st lattice
store a ! save this solution
write ex05.out1 ! write file for print-out
lch !
invert ! H -> n and n -> H
go ! again with N-refls only, for other lattice
loh ! list H refls for 2nd lattice
cell ! cell etc. for 2nd lattice
store b ! save this solution
write ex05.out2 ! write file for print-out
compare a b ! compare the two solutions
end !
The super lattice is geometrically correct, but we know better. The lattices with V=833.7 can be transformed to monoclinic C.
NOTE: as usually some reflections fit into both twin lattices. This example is also presented in the DirAx Course.

Example 6

ex06 - Inaccurate data. Start with default run and note that all solutions imply only a few fitting reflections. Raise IndexFit to 9 (for example). Now the best solution fits more or less 24 reflections.
dirax example 6 go ! inaccurate data, try default run first
loh ! only a few fitting reflections
lon ! note that many refl's could be made fitting...
indexfit 9 ! ...with less severe IndexFit
acl ! calculate solutions with new IndexFit
auto ! select default solution
lo ! note errors: about 10 times as large as normal
end !
Reflection nr 10 is a real outsider.

Example 7

ex07 - Not very accurate data and rather big(?) cell. A run with default Levelfit 1000 gives a solution with not-fitting reflections which might be halves.
Raising IndexFit does not help. Try a complete new run with LevelFit 300, now an all-fitting solution is found. Which solution is to be preferred crystallographically cannot be decided from this data alone.
dirax example 7 go ! default run
indexfit 4 ! try higher IndexFit
acl ! display all solutions
auto ! and select 'best' one
indexfit 2 ! restore default
levelfit 300 !
example 7 ! read data again
go !
lo ! show results
end !

Example 8

ex08 - Twinned crystal and ill-distributed data. The first default run gives a cell volume of 933.06. A second run with the 'n' reflections only gives a volume of 2806.8. Finally, a last run with the reflections that have become 'n' now gives a volume of 2799.8, so the problem is solved. This shows that you may have to handle DirAx cunningly.
dirax example 8 !
go ! first run
lo ! show results
store a ! save it
lch invert ! change n to H
go ! second run
store b ! save it
lch invert ! change new n to H
go ! third run
store c ! save it
compare b c ! compare the last two solutions
end !

Example 9

ex09 - Twinned crystal, giving a main lattice and a sub lattice. Default run. Select the default solution (volume 2863). Another cycle with 'n' reflections. Select default solution volume 5731).
dirax example 9 !
go !
cell ! accept acl=17
lch invert !
go !
cell ! accept acl=12
end !

Example 10

ex10 - Area detector data from a protein. Set Dmax = 300.
Result: all reflections fit, cell volume 1,561,000 approximately. (Cell can be transformed to cubic F.)
dirax example 10 !
dmax 300 ! change default
go !
end !

Example 11

ex11 - This used to be a really difficult problem for older dirax versions. Default run. One lattice with 11 fitting reflections, another with 10. Reflections 3,13,15 and 20 do not fit at all.
dirax example 11 !
go !
lch invert !
go !
end !

Example 12

ex12 - Twin, included because it is so nice. Default run. Two lattices are found: the most fitting is the 'main' lattice, V=1973.3, 17 fitting reflections (nr. 3 will fit with IndexFitFactor 3.00), the next best fitting lattice is the twin sister, V = 1974.12, 7 fitting reflections. This is an exceptional situation: mostly a run with 'n' reflections is needed to find the twin lattice, as in ex09. Note that in both lattices h and k are integer for all reflections.
dirax example 12 go !
lo ! show fitting and nonfitting reflections
compare 18 7 ! compare the two lattices
end !
Go to Dirax Main Page
Go to Dirax Commands
Go to Dirax Course