Exercise in Comparative Genomic Hybridization - Part 2

Written by: Peter Hallin

This exercise deals with the visualization of data from Comparative Genomic Hybridization (CGH) and comparison of mulitple genomes

Gene conservation : The BLASTatlas

In this part of the exercise, we will use the 'GeneWiz' software, developed here at CBS. It is designed to visualize genomic properties e.g. DNA structural parameters and base compositions. A recent Web Services based implementated called the BLASTatlas enables users outside of CBS to programmatically run the GeneWiz software and create an atlas visualization of any DNA sequence. You will use a pre-written Perl script which is a Web Services client, to visualize CGH data alongside with BLAST comparison of the reference against other strains.

GeneWiz reads numerical values - one per nucleotide position in the genome - and applies a user defined color scale and plots binned data on a circular representation. The BLASTatlas method goes a step further by allowing the user also to include maps showing the homology between the reference genome ('query') and any number of other genomes ('databases'). Each 'database' will appear in its own lane. We have downloaded the proteomes of two E. coli and one Shigella strains and the Web Services script will read these sequences and upload them to the BLASTatlas server.

For further reading, please consult the BLASTatlas homepage or see the reference below.


  1. BLASTatlas example: identifying conserved genomic regions

    2. Take a minute to look at this Clostridium BLASTatlas example. On page 2 of the PDF, you will find the legend.

QUESTION I: Familiarize yourself with the plot - try to identify one or two regions, where only the three C.botulinum strains display conservation on the protein level. That is, other Clostridum species should have no or little conservation.

You will now submit the three proteomes and CGH maps which contains data from STX negative, STX positive (STX, Shigella-like toxin), and E. coli EDL933 hybridizations.
  1. Submit your job to the BLASTatlas service

    2. This step will take roughly 2-4 minutes to execute. Log in to the computer system using SSH, having X11 enabled. Enter the directory for this exercise and run the Web Services script to produce a full view of the K12 genome. When the script has finished (2-4 minutes), view the output postscript.

    cp -r /home/projects/pfh/2006-03-01_PhD/exercises/cgh2 ~/
cd ~/cgh2
perl full.pl > full.ps
ghostview full.ps
QUESTION II: Do you see a general correlation between the BLAST lanes of the other O:157 strain (Sakai) and the reference strain (EDL933)?

QUESTION III: Is the correlation between the two O157 strains validated by CGH data?
  1. Zooming

    2. By editing the provided script (full.pl), you can zoom in on individual regions of interrest. We will now look for the Shigella-like toxins (stx1AB). You should not directly modify the full.pl script, rather you should make a copy and change that instead. Find the out-commented zoom specification at line 39-40 (you can enable line numbers in the menu: Preferences->Show Line Numbers). ALSO you must remove the out-commenting of the window specification at line 223 

    cp full.pl stx1.pl
nedit stx1.pl
perl stx1.pl > stx1.ps
ghostview stx1.ps

    GenesWindow
    stx2A, stx2B1352000 .. 1354000
    FlgX1573000 .. 1586000
    FlhX,MotAB2639000 .. 2656000
    FliX2696500 .. 2724000
    stx1A, stx1B2995000 .. 2997000
QUESTION IV: Repeat the method from above to produce a plot for stx2AB (see table below). Is the output as you would expect?

When determining the properties of a bacterial pathogen, not only the presence of toxins are relevant. Other features like motility and ability to secrete proteins are important. Below is a list of regions you may find interesting - feel free to browse around. To get a detailed list of the coordinates of the annotated features, you can view the genbank record of E. coli EDL993





LITERATURE
Hallin PF, Binnewies TT, Ussery DW: The genome BLASTatlas-a GeneWiz extension for visualization of whole-genome homology.Mol Biosyst 4:363-71 (2008) | Download PDF
This file was last modified Tuesday 13th 2008f May 2008 15:35:12 GMT