Some of the known REPs did not come up in any of the searches, and
others rarely appeared.
Table 8 
on page
lists the REP sequences that were rarely found.
Most of the missed sequences are fairly far from the consensus
sequence, but three of the misses are rather surprising: REP39- (which
is found only by REP99-gx6), REP60 (which is found only by
REP99-gxn-fromseq14, REP99-gxn-fromseq15, and REP99-gxn-fromseq16),
and REP83 (which is found by most of the
HMMs, but not by the simple Markov models).
Table 8: REP sequences that were not found by most searches.
The first column lists the number of searches (out of 81) that failed
to find the REP. The name is from the list compiled by
Rudd [13].