(1) Start with a set of sequences that have a common motif. This works better if you have a large number of samples.
(2) Guess the length of the motif, and guess the location of the motifs in each sequence, or guess the motif. Prior information helps! (e.g. DNA footprint analysis)
(3) Assign a probability score to each possible location of the motif in each dataset.
(4) Refine the guess of the motif based on the probabilities. Repeat steps (3) and (4) until convergence.
MEME
| Previous slide | Next slide | Back to first slide | View graphic version |