 |
|
| TITLE: Finding Motifs Computationally |
|
|
ABSTRACT:
Finding common patterns, or motifs, in the promoter regions of
co-expressed genes is an important problem in bioinformatics.
Despite years of efforts in solving this problem, less than
30% of the known motifs in TRASNSFAC can be found
computationally.
Motifs can be represented by strings or probability matrix.
In this talk, we shall present approaches and formulations
of this problem based on these two common motif representations.
The planted (l,d)-motif problem (PMP) is formulated for string
representation, where l is the length of the motif and
d is the maximum Hamming Distance between the similar patterns.
We shall introduce new algorithms to solve this motif
problem and its variations.
Finding motifs based on matrix representation is very
difficult. Even for a motif length 6 or 7, there is no
algorithm that can guarantee finding the exact optimal
matrix from an infinite number of possible matrices.
New advances in finding the optimal matrix will be discussed
in this talk.
Finally, some new motif discovering algorithms, which assume
extra biological knowledge and other types of representations
to greatly improve performance, will be highlighted. |
|
|
