Volume Visualization using Chimera

Tom Goddard
November 8, 2001
NCRR site visit

All images on this page ©2004 The Regents, University of California; all rights reserved.

Introduction

• I've been developing volume display in Chimera to assist several research projects.
• Volume data means 3D arrays of numbers
• Examples: Electron microscope, light microscope, crystallographic electron density maps, ...
• Will show 2 Chimera extensions that provide display and path tracing of volume data.
• Main message of this talk: Chimera is useful for analyzing data over a wide range of resolutions.

Large Range of Data Resolutions

Crystallographic density

Electrostatic potential

Water occupancy map (1A)

Electron cryo-microscopy (7A)

Electron microscopy (300A)

Light microscopy (2000A)

Barcoded Chromosomes

Arrangement of chromosomes in the cell nucleus. Collaboration with Mike Lowenstein in John Sedat's lab, UCSF. 3D light microscope images of chromosome 2L in Drosophila nuclei. Can determine whether specific stretches of DNA are attached to the nuclear envelope. Can study spatial arrangements characteristic of differentiated cell types. Can study relation between gene transcription and spatial arrangement.

Tracing Chromosomes

Drosophila chromosome 2L has 13 fluorescently labeled segments using 3 fluorophors. Trace path of chromosome by finding expected sequence of color spots in light microscope data. Different colors provide improved resolution Colors allow correct interpretation when spots are missing Prior to Chimera, tracing was done by flipping through a stack of 2 dimensional images and typing observed sequence of spot id numbers into a file. DAPI stain in light microscope images illuminates all DNA. Nuclear envelope can be shown as boundary of DAPI stained region.

Low Resolution Protein Structures from
Electron Cryo-Microscopy

Rice dwarf virus capsid protein structure determined 7 angstrom resolution is sufficient to model alpha helices and beta sheets. Single particle cryo-EM can handle large complexes and does not require crystalization. Useful for determining protein folds Can determine relative orientations of components of complexes Work is done by Wah Chiu's group at Baylor College of Medicine, National Center for Macromolecular Imaging. Alpha helices are found in density map by program helixhunter Helixhunter developed by Wen Jiang, Matthew Baker, in Wah Chiu's lab. Display predicted helices superimposed with EM data.

Tracing Connections between Helices

Interactively trace turns connecting helices. Goal is to find correct helix order, not exact turn structure. Helices predicted from amino acid sequence can be matched to those found in density map. Without Chimera, Wah's group did turn tracing on 2 dimensional images in Photoshop. Rice dwarf virus analysis was completed 5 months ago at start of our collaboration. Chimera not used for results published in October issue of Nature Structural Biology. Homology to known crystal structure of Bluetongue virus P7 capsid protein used in tracing turns.

Strengths of Chimera Volume Visualization

• Molecular models, VRML for low resolution models, and volume data
• Show volumes as transparent solids, meshes and surfaces.
• Marker placement and path modeling
• Precomputed subsampling for large data sets
• Subregion selection with mouse, zone selections around markers and atoms
• Python extensions for custom modeling
• 3D image processing, ridge finding, gaussian filtering, ...

Future Development

• Multi-scale models
• Volume movie player with analysis capabilities
• Use Phantom 3D force feedback input device for marker placement and path tracing.
• Flood fill with interactive range control to select or sculpt away volume regions.
• Non-orthogonal data display for crystallography, accomodate skewing matrix
• Automated marker placement and path tracing.
• Interactive construction of triangulated surfaces using placed markers

How to Depict Macromodels

• Ribbon representation of large ribosomal subunit
• Coarser representations are useful for:
  • selecting components of interest
  • examining component assembly
  • moving structure interactively

Analyzing Macromodels: An Example

• How are nucleosomes arranged in the chromatin 30 nm fiber?
• Model of irregular nucleosome placement could come from electron microscopy. (Image shows helical placement with randomized orientations.)
• We have crystal structures of a nucleosome, ~150 bases DNA wrapped around histone core
• Show segment of fiber as aspirin tablets representing nucleosomes
• Limit view to a few nucleosomes and show atomic resolution detail
• Look at how histone tails form inter-nucleosome contacts

Macromodel Development Objectives

• Goal is to provide interactive visualization of models over a wide range of resolutions.
• Provide low resolution surface representation of macromodel components
• Support data formats for macromodels
• Target systems where we have atomic resolution structures
• Useful for virus capsids, actin / myosin systems, ion channels, ribosome complexes, collagen bundles, cyto-skeleton, nuclear matrix, ...

Volume Time Series

• Provide volume movie player with analysis capabilities
• Examples from current collaborations:
  • Understanding water localization around collagen
  • Tracking chromosome transformations during the cell cycle

Water Dynamics around Collagen

• Collagen water localization changes over the course of molecular dynamics run
• Do preferred water positions vary as collagen bends?
• Are apparent water positions statistical artifacts?
• Look at water occupancy volume movie to guide quantitative analysis.
• Allow comparing 2 or more sections of volume movie simultaneously with sliders controlling position in movie.

Chromosome Dynamics during the Cell Cycle

• John Sedat's group has made movies of fluorescently labeled chromosomes in live cells.
• Track motion of labeled segment of chromosome during cell cycle.
• Localization in certain areas may correlate with gene transcription.
• Manual or automatic spot tracing capabilities needed.