mrc software for 2d crystal structure determination

MRC software for 2d crystal structure determination

Henderson & Unwin (1975) Nature 257, 28-32

• Introduction

• Basic Functions and Principles

History Developed over 30 yrs by a number of people Written in Fortran image2010

image format compatible with ccp4 plots in postscript format

http://www.tonh.net/museum/plotter.html

Documentation Crowther, Henderson and Smith (1996)

J. Struct. Biol.116, 9-16 Docs in the package: image2010/doc/ Docs on 2dx.org Help Docs in 2dx applications

C LATLINPRESCAL : program to correct image amplitudes for effects of CTF andC to calculate appropriate weights for output to LATLINE. C******************************************************************************** C Control cards: C C Card 1: NSER,ZMIN,ZMAX (*) C serial number on ORIGTILT input C and ZSTAR limits to be passed on to LATLINE. C Card 2: IQMAX (*) C maximum value of IQ for spots to be used C C INPUT file is raw file straight from ORIGTILT : on unit FOR001 C OUTPUT has the format required by LATLINE : on unit FOR003 C C*******************************************************************************

3D reconstruction from 2D TEM images

Unwin, Scientific America

Single Image Processing

Unbending

Boxing MMBOX

Fourier Transform

Unbending is possiblebecause distortion can be tracked along the lattices

The distortion is plotted at 10x the actual values

QUADSERCHK

Reference from rough 3D reconstruction

MAKETRAN

Reference from best region in the image

Cross Correlation

Correlation against perfect crystalline structure gives distortion map

Determine Amplitudes and Phases of Diffraction spots MMBOX / TTBOX + CTFAPPLY

h k Amp Phase IQ Background0 1 434.9 34.2 3 156.8 0 2 7924.5 44.6 1 148.50 3 1373.7 210.6 1 158.30 4 1411.4 228.4 1 124.30 5 710.5 206.4 2 106.30 6 1870.8 126.0 1 121.1 .

.

.

Amos, Henderson & Unwin (1982) Prog. Biophys. molec. Biol., 39, 183-231

IQ values determined by

amplitudes Used as measure of

phase error

Relationship between Amplitude and Phase Error

Re

Im

IQ Amp/(RMS background) cutoff

Phase Error

1 8.18 7

2 4.09 14

3 2.73 21

4 2.05 28

5 1.64 35

6 1.36 42

7 1.17 49

8 1.02 56

image2000/source/mmboxa.for

IQ~4

IQ~2

Merging information and Refinement

Unwin, Scientific America

ORIGTILTLATLINE

ORIGTILT program Move the projection images (i.e., phase origins) to a

common center and compatible orientation

Reference Symmetry special position Last set of merged image Last calculated 3D model

Refine Phase origin Tilt geometry Beam Tilt

Standardize a, b axes Rotation Flip crystal upside

down Defocus inversion

Lattice line fitting (LATLINEK) model fitting in the reciprocal space of the continuous

line in the non-repeating (z*) direction based on the assumption that the density in the real space is limited by a window function

Agard (1983) JMB 167,849-852a=D

=

D

1

a

Stahlberg

LATLINEK Parameters

RMIN RMAX

RCUT

PFACTprotein density

ALAT = c in 3D xtal lattice parameterDELPRO

BINSIZE

D=RMAX-RMIN

GUESS GENERATE CURVE, FIT STATS, and FIT SAMPLE

SAMPLE CURVE for PLOT

REAL SPACE

1/D DELPLT

1/ALATFOURIER SPACE

AK, IWF, IWP for weighting of amp error vs. phase error of the fit.

Importance of Defining Correct Window Function Width

D=260 A D=100 A

y=a0+a1xy=a0+a1x+a2x2

Avoid mistakes Unbending

Reference should look like perfect crystalline structure

Good Bad

Exacting Phase/Amplitude Real signal should become

significant when added up

Merging/Refinement Projected symmetry should be

obvious in untilted image and its Fourier transform

Choose window function width according to property of the specimen

8 IQ8 spot does not make the result as reliable as 1 IQ1 spot

8-6 A7 8 9 9 87 10 13 11 78 8 11 11 87 7 10 11 97 7 9 9 8

5-3.5 A

7 7 7 7 78 7 6 7 77 7 7 7 67 8 7 7 77 7 7 7 7

5 Å

7 Å10 Å

What MRC package can not do Make your first crystal Merging results from different crystal form Untwist Rotational disorder in the crystal Correct for bending perpendicular to the

membrane Merging information from unknown and variable

number layers of stacked crystals ….

Take this home: Computers have

changed but the principle is the same.

Keep a critical eye during the process while enjoying the express ride of 2dx and IPLT.

A 1947 or prior photograph (cropped) of IBM (Hollerith) operators at the Caltech Computing Center machine room performing crystallographic computation using the IBM card punch systems as described in Shaffer et al. (1946) and briefly summarized in Swift (1947).

http://journals.iucr.org/a/issues/2008/01/00/sc5015/sc5015bdy.html