RECON -- finding repeat families from biological sequences

This repository contains a development version of

of RECON. The last official release was initially

checked in with minor documentation changes as

Commit def1f6f. Currrently the master branch contains

code modifications that are in-flux. When the 3.0.9

release is ready we will begin to keep the master

branch synced with official releases.

Copyright (C) 2001 Washington University School of Medicine Copyright (C) 2011-2021 Institute for Systems Biology

About the package

The RECON package implements a de novo algorithm for the indentification of repeat families from biological sequences. For details about the algorithm, see

Bao Z. and Eddy S.R. (2002) Automated de novo Identification of Repeat Sequence Families in Sequenced Genomes. Submitted.

and

http://www.genetics.wustl.edu/eddy/recon

Source code of RECON is freely available under the GNU General Public License. For details on the copyright and licensing, see the COPYRIGHT and LICENSE file.

This distribution includes two Perl scripts and several C programs. Most users only need to run the scripts and the C programs should be transparent.

What's new?

RECON1.08: Wed Feb 5 14:04:25 PST 2014 - Robert Hubley ISB

• New checks in eleredef.c for failed memory allocation/file operations.
• A buffer overrun bug fix provided by Stephen Ficklin in eledef.c.

RECON1.07: Wed Jun 8 10:00:55 PDT 2011 - Robert Hubley ISB

• Another problem with the eleredef program has appeared a couple of times. A divide by zero error in one of Zhirong's routines. I have applied a workaround to the divide by zero but not a fix to the real problem. The program will now warn the user if the workaround was exercised during a run.

RECON1.06: Tue May 20 15:19:30 PDT 2008 - Robert Hubley ISB

• 64 bit compiler fixes: Previous versions of RECON used the "long" data type under the assumption that was 32bits. This caused the program to crash or lock up during analysis.

Bug fix for RECON1.02 in the eleredef program.

Programs and input

The RECON algorithm defines repeat families on the basis of pairwise similarity between the input biological sequences. We do not provide pairwise sequence comparison tools in the package. However, we provide a tool (MSPCollect) which converts a BLAST output into an MSP file, which is used by the package as input. In the MSP file, each MSP that BLAST reports is turned into a one-line summary in the following format:

score %iden start end query_name start end subject_name \n

The command line format for MSPCollect is

MSPCollect BLAST_output_file > name_of_your_MSP_file

If you have multiple BLAST output files, you have to run MSPCollect multiple times and concatenate the result into one MSP file.

The recon script drives the running of the package. The command line format is as follows:

recon.pl seq_name_list_file MSP_file integer

The seq_name_list_file contains the names of the input sequences, one per line. The lines can be just the names, or it may be in the FASTA format (starting with ">"). The names should be sorted in lexical order. The very first line of file should be an integer which is the number of names listed in the file. If the input sequences are stored in one FASTA file, one can construct the seq_name_list_file as follows:

grep ">" FASTA_file | sort > name_of_your_seq_name_list_file

One can then add the number of names to the top of the file.

As one of the early steps, the script sort the MSPs. If there are too many MSPs in the input, one may consider sorting a subsection at a time. The third integer argument specifies how many sections you want to divide the sorting process into. The default is 1. (Notice, as the file size gets big, the sorting process becomes inefficient, and may even cause the process to crash.)

If you know the RECON algorithm, then the names of C programs are self explanatory.

Output of the package

The package generates several directories along the way. All of them, except for the summary directory, stores intermediate results. The ele_def_res, ele_redef_res and edge_redef_res directories contain files for the elements defined at the corresponding step, one file per element. Elements are indexed using integers and are named as e123, e987, etc.. If you are not interested in the details about the elements, you can remove all these directories.

Two files in the summary directory are the "official" output of the package -- eles and families. Each line in "eles" is a summary of an element defined in the following format:

family_index element_index strand sequence_name start end

Each line in "families" is a summary of a family defined in the following format:

family_index copy_number_in_the_family

The Demo

To get examples of (the format of) the input and output files, see the README file in the Demos directory.