Thematic image University of Washington Department of Computer Science & Engineering
 Frequently Asked Questions on YMF
  CSE Home    YMF Home   Send Mail  Download 

Where can I find all the details about YMF ?

The algorithm is described in :

The user interface for the web server is described in : The results of applying YMF on various data sets from yeast are in:

How does YMF compare to other motif finding programs ?

There is no clear-cut answer to that, since there is no universal benchmark for testing motif finding programs.We have compared the performance of YMF with MEME and AlignACE, and the results can be found in:

YMF evaluates every motif in its search space for its over-representation, and is thus guaranteed to report the best motifs by its scoring criterion. This is in contrast to the local search techniques deployed by most other motif finding programs. YMF can afford this enumerative search because its motif model is simpler than the weight-matrix model used by other programs. YMF's motif model was derived from an empirical observation of several biologically determined transcription factor binding sites in yeast, and the variations in the sites bound by the same factors.

In our comparison experiments on yeast regulons for which the motif was known, we observed that YMF performed better than MEME and AlignACE, more often than not. When the known motif conformed to its model, YMF performed better, as expected. However, there are cases where the known motif cannot be reasonably captured by YMF's motif model, and in such cases, MEME or AlignACE did better. Hence, we suggest using some other motif finding algorithms in conjunction to YMF, for greater sensitivity.

What are some other motif finding programs on the Web ?

You may want to try out (all of the following find weight matrix motifs):

Does YMF search for TRANSFAC motifs ?

No, YMF is intended to discover novel binding sites. It was not designed to find instances of known motifs (from TRANSFAC or SCPD or any other motif database) in input sequences. It may happen that a motif discovered by YMF is very similar to a motif from one of the databases, but that will be because the motif is over-represented in the input sequences, and not because it is present in a database.

Why does the YMF form ask me for the organism name ?

YMF evaulates a motif by comparing its observed number of occurrences (in the input sequences) to the number of occurrences expected in a random sequence representative of promoter regions in the organism under study. If the composition of the promoter sequences changes significantly from one organism to another, the evaluation of the motif will also change. Hence, YMF requires knowledge of the organism under study, in order to make an accurate evaluation of motif significance.

What if my organism is not in the list given in the form ?

You have two choices:

I successfully ran YMF on my sequence. The "Get output in text format" link gives me a very large number of motifs !!

Correct, and if you look closely, most of the motifs are very similar to each other. This is expected. YMF outputs the top 1000 motifs (that fit the characterization given by the input parameters) sorted by their significance scores. It is normal that this list will have many minor variations of the same motif. These are statistical artifacts of the truly significant motifs. If you are just interested in the top motif, use option (2) "Add the top motif to the session". If you want to extract the top few motifs from this list, that are not variants of each other, use option (3) "Run FindExplanators".

Can I use masked sequences as input ?

Yes, you can, as long as the masking character is either 'N' or 'X'.

I am having problems in plotting my motifs.

As of now, the plotting feature has been tested only on Windows machines using IE6.0+.

My question is not here.

Please send e-mail to We will be happy to address your concerns.

Close this window