From: benson@ecology.biomath.mssm.edu
To: "@msvax.mssm.edu":@msvax.mssm.edu:mgreese@lbl.gov
Cc: benson@ecology.biomath.mssm.edu
Subject: Re: Information on submissions
Date: Fri, 23 Jul 1999 15:53:25 -0400 (EDT)

Dear Martin,

	Below is the additional information you requested for my
submission to the compfly project for ISMB99.  Unfortunately, I will not
be able to attend the tutorial.  I will be there on the 9th, 10th and
11th and give a talk on the afternoon of the 10th.  I hope we will have
time to talk about the results.  

				Best regards,

				Gary Benson

---------------------------------------------------------------------
Tandem Repeats Finder Program  --  Compfly Annotation for ISMB99

1. Gary Benson
   Department of Biomathematical Sciences
   Box 1023
   The Mount Sinai School of Medicine
   One Gustave L. Levy Place
   New York, NY 10029-6574

   (212) 241-5777 work
   (212) 860-4630 fax
   email: benson@ecology.biomath.mssm.edu
   url: www.mssm.edu/biomath/benson.html


2. Annotation obtained by running Tandem Repeats Finder v2.02 on the
entire 2.9 Mb file using default parameters.  Run time was 40 seconds.  
Output from Tandem Repeats Finder converted to GFF format with a perl 
program.  No manual intervention.  

3. Tandem Repeats Finder locates approximate tandem repeats in
nucleotide sequences.  A tandem repeat is two or more contiguous,
approximate copies of a pattern of nucleotides.  The program is able to
identify tandem repeats without the need to specify either the pattern
or the pattern size. Repeats with pattern size from 1 to 500 bases
are detected.  The program uses a theoretical model of tandem repeats
which specifies how much similarity must exist between adjacent copies. 
For example, the user could specify that the copies be, on average, at
least 80% similar.  Additionally, an average number of indels between
corresponding positions in adjacent copies is specified, for example
10%. Based on the statistical model, we developed a number of
statistical criteria for detection based on distributions of head runs
in Bernoulli trials (coin tosses) where the probability of heads is the
same as the average similarity.  

The program uses k-tuple matches (short word matches) to locate
candidate tandem repeats.  The required number of matches and their
overall separation are based on the statistical distributions. 
Alignment by wraparound dynamic programming is used to validate the
candidates.  

Output from the program consists of at least two html files which can be
viewed through a web browser.  One is a summary table of the repeats
detected, including their location, pattern size, number of copies and
nucleotide content.  This information was converted to the GFF format. 
A second file contains the alignment of each tandem repeat against its
consensus sequence.  The summary table and alignment file are linked so
that clicking on an entry in the summary table brings up the correct
alignment.  The output files from this annotation project can be viewed
at http://c3.biomath.mssm.edu/trf/Adh.fa.2.7.7.80.10.50.500.1.html on
our webserver.  

4. Reference:
G. Benson,
"Tandem repeats finder: a program to analyze DNA sequences"
Nucleic Acids Research (1999)
Vol. 27, No. 2, pp. 573-580.

5. Website:
http://c3.biomath.mssm.edu/trf.html

Tandem Repeats Finder is available for use over the internet and 
Unix or PC versions can be downloaded for local use.