A set of fixed-length sequences located at the same relative position
with respect to a functional site (eg. a transcription start site)
is scanned for occurrences of a given motif. The motif may be defined
by a consensus sequence or a weight matrix. The search may be confined
to a sub-region within the sequences supplied. The program returns either
the sequences containing the motif, the sequences not containing the
motif, or the sequence regions around the motif.
See the here