DIRDIF ROUTINE STRUCTURE SOLUTION USING YOUR CHEMICAL KNOWLEDGE

1
DIRDIFROUTINE STRUCTURE SOLUTION USINGYOUR
CHEMICAL KNOWLEDGE

• Paul T. Beurskens
• Gezina Beurskens
• Rene de Gelder
• Santiago Garcia-Granda
• Robert O. Gould
• Jan M.M. Smits
• and others
• 
  
  Marieken van Nieumeghen
• DIRDIF

2
DIRDIF DEMO NANCY 2006 INTRO
4 vector search
ROT POSN model
3 H.A. Patterson
PATTY
ORIENT
TRACOR
1 DIR meth. gt DIF struct.
TRAVEC
PHASEX
2 R2 AUTO recycling
5 TRICKS and HELP
FOUR
6 WHAT is NEXT?
summary
3
INTRO DIRDIF LOGISTICS
organizer
DDSTART
supervisor
DDMAIN
PATTY
ORIENT
TRACOR
TRAVEC
PHASEX
R2
FOUR
utilities
NUTS
DIRDIF finished
4
MONOS DIRDIF TEST STR.
DDSTART
MONOS DEMO 1 full structure DEMO 2
S only DEMO 3 S misplaced
0.125A centric ! run DD
FOUR NORECY

DDMAIN
FOUR
NUTS
DIRDIF finished
5
AB INITIO DIRECT METHODS

• the problem starting
  phases?
• sigma-2 or
  result
• tangent formula

6
DIRECT METHODSapplied to DIFFERENCE STRUCTURE
FACTORS

Fp part.str.f. known Fr part.str.f. unnkown
7
DIRDIF
DDSTART

DDMAIN
MONOS DEMO 4 S
only with PHASEX run DD PHASEX
NORECY
PHASEX
FOUR
NUTS
DIRDIF finished
8
program PHASEX

• delta F conventional Fourier coeff.
• starting for tangent formula
• Fr PHASEX-refined str.f. of
• the remainder of the structure
• enantiomorph problem

symbolic addition with 999 known phases 10
unknown lettersymbols
9
ENANTIOMORPH PROBLEM
DDSTART

DDMAIN
MONOS DEMO 5
S misplaced
0.135 A centric! run DD
PHASEX NORECY repeat!
PHASEX
FOUR
NUTS
DIRDIF finished
10
H.A. PSEUDOSYMMETRY

• SbS4.....

11
1963 DIRECTE METHODEN

12
advantage of PHASEX direct methods

• the known part of the structure
• is subtracted from the observed data

• many phases known strong starting set
• number of atoms reduced
• heavy atoms subtracted
• the cause of problems in conventional direct
  methods
• is subtracted from the structure

• -- heavy Patterson overlap usually comes from a
  fragment of the molecule that can be
  used as a partial structure
• -- strong pseudo-symmetry is subtracted

13
2 R2 AUTO recycling
4 vector search
ROT POSN model
3 H.A. Patterson
PATTY
ORIENT
TRACOR
TRAVEC
PHASEX
2 R2 AUTO recycling
5 TRICKS and HELP
FOUR
6 WHAT is NEXT?
summary
14
R2-driven AUTOMATIC RECYCLING

• remove one atom
• R2 lower? atom gone!
• R2 higher? get it back!

1969 DUAL SPACE recycling
direct space electron density
interpreation chemical knowledge cell contents,
H.A., geometry
reciprocal space direct methods R2-criterion
MONOS DEMO 6 S only 6a correct 6b
misplaced run DD PHASEX
15
partial structure expansion by DIRDIF
CAROL
3 mol/cell
DDSTART
DDMAIN
CAROL DEMO DD FOUR NORECY not
good DD PHASEX NORECY DD ,,
,, DD ,, ,, going
better DD PHASEX perfect
PHASEX
FOUR
NUTS
DIRDIF finished
16
Patterson programs in DIRDIF
PATTY
ORIENT
heavy atoms
molecular fragments
TRACOR
TRAVEC
R2
DIRDIF finished
17
3 H.A. Patterson
4 vector search
ROT POSN model
3 H.A. Patterson
PATTY
ORIENT
TRACOR
TRAVEC
PHASEX
5 TRICKS and HELP
FOUR
6 WHAT is NEXT?
summary
18
HEAVY ATOM PATTERSON

• PROGRAM PATTY
• solves crystal structures faster
• than the chemists
• can make crystals
• FOM lt P (vij ) gt
• vij rj(s) ri(s)
• i, j atom
• s, s symmetry
• r tentative atomic position from the minimum
  function
• v interatomic vector (based on tentative
  positions)
• P observed Patterson (function value for a
  vector)
• FOM figure of merit using Nordman weights

19

heavy atom structures with DIRDIF
DDSTART

DDMAIN
DEMO MONOS run DD PATTY
PATTY
PHASEX
FOUR
DEMO KAP2 run DD PATTY
NUTS
DIRDIF finished
20
origin problem (superstructure)
NGUY
origin and enantiomorph
PBAG

• symbolic addition on REAL part
• symbolic addition on IMAGINAIRY part

21
OXATON - what to do ?

• 3 independent molecules
• 3 S O5 C19
• PATTY S ?
• ORIENT SC7 ?
• or OC8 ?

22
vector search And rec. space correlation
functions 4
4 vector search
ROT POSN model
ORIENT
TRACOR
TRAVEC
PHASEX
5 TRICKS and HELP
FOUR
6 WHAT is NEXT?
summary
23

vector search with DIRDIF
(r)
(r)
(r)
ORIENT
TRACOR
TRAVEC

• t

DIRDIF finished
24
ORBASE and ORUSER

• ACKNORT

WRONG MODEL ?
25
MODELS FROM YOUR OWN WORK !
JOOST1
flexible
model BASKET

• 2007
• VECTOR SEARCH AS FIT AS A FIDDLE

26
USING THE CAMBR. CRYST. DATA BASE
1976 VECTOR SEARCH
MUEN

• HUMMEL

?
!
NOT OK USE NEXT MODEL
27
5 TRICKS and HELP
5 TRICKS and HELP
6 WHAT is NEXT?
summary
28
6 WHAT is NEXT?
6 WHAT is NEXT?
summary
29
MULTISOLUTIONS

• EB2207
• Combine PATTY and R2 results
• QFOM SQRT( PFOM R2expect ) / R2
• PATTY-nr PFOM R2exp R2 QFOM
  best output
• 1 1.000 .422 .616 5
  1.055 4 - PFOM
• OK 2 .948 .422 .626 6 1.010
  5
• OK 3 .913 .422 .557 2 1.114
  2
• OK 4 .891 .562 .608 3 1.163
  1 QFOM yes
• 5 .866 .422 .610 4
  0.991 6 -
• OK 6 .842 .422 .556 1 1.072
  3 R2
• 7 .795 .422 .769 7
  0.753 7 -
• MARS

for PATTY TRACOR ORIENT similar
30
TRANSLATION SEARCH

• some peaks
• must be wrong

31
DIRDIF DEMO NANCY 2006
PATTY
ORIENT
TRACOR
TRAVEC
PHASEX
R2
FOUR
summary
32
DIRDIF DEMO NANCY 2006
Thank you for your attention
PTB
33
DIRDIF DEMO NANCY 2006
DIRDIF unified 2006
34

• marieken300
• 300x300 dpi
• 122 kb

35
archief

• kristalman.jpg