|
Motif Discovery Assessment: Summary |
|
 CSE Home |
About Us |
Search |
Contact Info |
|
|
The motivation for the assessment is that there are numerous motif
discovery programs available, but biologists have almost no guidance to
help them choose which one to use. The reason for the
organizer not to run all the motif discovery programs unaided is that
many of these programs require experienced setting of several
parameters. We would like those choices to be made as well as
possible, so that the assessment compares the programs running at
their most accurate performance. A second motivation for the
assessment is to construct a benchmark of data sets for assessing
future motif discovery algorithms.
We have submitted a journal paper describing the assessment and its results. The journal paper and assessment website have pointers to the participating algorithms and software, as well as information on how participants chose parameters, any pre- and post-processing, etc. The assessment participants completed their predictions in February 2004. In a nutshell, the organizers supplied data sets, each consisting of a collection of DNA sequences in which we know the locations of planted binding sites. Participants ran their own programs on each data set and sent the organizers the top motif prediction for each data set. The organizers then compared the predictions to the known binding sites and reported the results. You can find more details on each of these steps by following the links labeled "Next" in the navigation bars on the left of each page. |
 |
Computer Science & Engineering University of Washington Box 352350 Seattle, WA 98195-2350 (206) 543-1695 voice, (206) 543-2969 FAX [comments to tompa] |
