START Version 1.0.8

Written by Keith Jolley.
BURST code by Ed Feil and Man-Suen Chan.
Index of Association and Homoplasy Test implementation modified from code provided by John Maynard-Smith.

START is provided on an "as is" basis, and the authors expressly disclaim any and all warranties, express or implied, including, without limitation, warranties of merchantability and fitness for any particular purpose. In no event shall the authors be liable to any party for any direct, indirect, incidental, punitive, or consequential damages of any kind whatsoever with respect to use of this product.

Copyright 2000-2001 University of Oxford, All rights reserved.

START (Sequence Type Analysis and Recombinational Tests) is a selection of tools useful in the analysis of MLST data. These tools can be broadly divided into four categories: 1) data summary, 2) lineage assignment, 3) tests for recombination and 4) tests for selection. This is very much a work-in-progress and it is envisaged that more analyses will be added to the package over time.

This program may be used and distributed freely but only as part of its original installation package. We would be grateful for any comments, suggestions or bug reports. Please send these to

The latest version of this software is available for download from

System Requirements

This software is for systems running Microsoft Windows 32 bit operating systems. It has been tested on Win 95, 98, NT, 2000 and XP.

The software utilises an embedded web browser control for data output and consequently requires Microsoft Internet Explorer 4.0 or higher to be installed. If this is not installed, an error reporting "component SHDOCVW.DLL or one of its dependencies not correctly registered: a file is missing or invalid" will be displayed. If required, this may be downloaded from

Windows2000 and XP users: START utilises a configuration file within the program installation directory, usually 'Program Files\START'.  If your system security profile is not set to allow user modification of this directory, an error message concerning 'Form/Load' will appear when trying to run the program. Changing the permissions will resolve this problem.  If you do not know how to change the security settings on this directory, please consult your system administrator.

The BURST algorithm creates postscript files for some analyses. In order to view these, we recommend installing Ghostscript and Ghostview. If you associate .ps files with Ghostview, the application will display postscript output when you click the appropriate hyperlink in the results display window. Ghostscript and Ghostview are available free for download from

Known problems

The printing funcitonality used to work by pressing the print button or selecting 'print' from the 'File' menu. This seems to have stopped working with some of the latest Microsoft updates. You can still print from START by right-clicking on the output window and selecting 'print' from the context menu that appears.

Revision history

Version 1.0.8:

Fixed overflow error in the homoplasy test code that caused the program to crash under certain circumstances.  This version is available as a patch.  Simply download the start.exe file and replace the previous version.

Version 1.0.7:

Fixed overflow error that caused the program to crash when calculating dN/dS ratios when there were more than ~256 alleles loaded for any locus.

Version 1.0.6:

Fixed bug in the Maximum Chi Squared implementation that over-calculated the significance of some recombination events.  This was mainly a problem when looking at two parental sequences and a derived.

Version 1.0.5:

Introduced check for invalid bases in sequence files.

Fixed bug that resulted in a divide-by-zero error when calculating dN/dS ratios when only two sequences were being compared or when there were no silent polymorphic sites.

Fixed bug in Sawyer's module that caused the program to crash when there were no silent polymorphic sites.

Version 1.0.4:

Fixed bug in the Maximum Chi Squared implementation that under-calculated the significance value of recombination events.  The putative site of recombination was correctly identified.

Version 1.0.3:

Fixed bug in the polymorphism frequency display that caused a crash when used with allele sequences greater than 1000 bases in length.

Contains new version of the BURST module that allows BURST analysis with datasets containing up to 2500 sequence types.  Previous versions were limited to 1000 sequence types.

Version 1.0.2:

Contains updated version of BURST module.

Version 1.0.1:

Fixed bug that produced an overflow error while trying to draw UPGMA trees containing more than about 1600 isolates.

Fixed bug in the parsing of the BURST output file. This had resulted in the mis-reporting of the number of satellite STs and occassional strange results.

Version 1.0.0:

The program now checks for the presence of spaces within FASTA sequences.

Revised help files.

Version 0.9.2 (limited release) Beta:

Data summary analyses have been significantly speeded up by caching the output file. Previously this was built up a line at a time and if a user had real-time virus checking enabled, this file would be tested each time it changed (possibly hundreds of times for some outputs).

Fixed bug that caused the program to crash during BURST analysis if the results contained no singletons.

Version 0.9.1 Beta:

Fixed bugs that caused the program to crash when clicking loci re-ordering buttons in the species dialog if no loci names had been entered.

Fixed bug that caused the program to crash when trying to load sequence files if no locus names had been set in the species dialog.

Fixed bug that caused the program to crash if fewer than seven loci were used for the Homoplasy Test.

Version 0.9.0 Beta:

First release version.