Looking for REPs in EcoSeq6

As an example of the techniques described in the preceding sections, the cost map method was used to find clusters of repetitive extragenic palindromic sequences (REPs) in the 1875932 bases of the EcoSeq6 database [12]. The sequences found were compared with a list maintained by Ken Rudd [13]. The three search techniques used for building this comparison list were described and referenced in Table I of [2]. The best of the techniques mentioned there (self-BLAST) found 106 of 112 REP clusters in EcoSeq5, or about 95%.

One goal is to do at least as well as the self-BLAST search in finding the already known REPs, and, hopefully, to provide a better characterization of the structure of the REPs than current consensus sequences.

The current consensus sequence for a REP [2] is

5'GCCKGATG-CGRCGY---RCGYCTTATCMGGCCTAC3'
where K is G or T, R is G or A, M is C or A, and Y is C or T.

A variant on the REPs, named REPv, has also been identified, and given the following consensus sequence [13]:

GCCTGATCGCGCTACGCTTATCAGGCCTAC.

• Looking for REPs using a seed
  • Using simple Markov models
  • Using hidden Markov models
• Looking for repeated elements without a seed
• Hard-to-find REPs