Position weight matrices (PWMs) are widely used to
depict the DNA binding preferences of transcription factors (TFs) in
computational biology and regulatory genomics. Thus, leaning an
accurate PWM to characterize the binding sites of a TF is a fundamental
problem that plays an important role in modeling regulatory motifs and
discovering the binding targets of TFs. We believe that the accurate
PWM of a TF shall yield the maximum likelihood of simultatneously
observing both its binding sequences
and the associated gene expression or ChIP-chip values. We
developed a new approach to learning PWMs via a
weighting scheme. The new learning approach has been
incorporated into the popular motif finding program AlignACE,
and the motified program is called W-AlignACE. The large-scale tests
have demonsrated that W-AlignACE is an effective tool for discovering
TF binding sites from gene expression or ChIP-chip data and, in
particular, has the ability to find very weak motifs.
Due to the stochastic nature of Gibbs sampling,
program W-AlignACE ( as well as AlignACE) may produce different outputs
in different runs. This issue still exits even when running W-AlignACE
with the same random seed but in the different platforms (because
different random generators may be employed).
Run W-AlignACE ::
- X. Chen and T. Jiang. An improved Gibbs
sampling method for motif discovery via sequence weighting. Proc. of Computational System
Bioinformatics, 239-247, 2006.
- X. Chen, L. Guo, Z. Fan, and T. Jiang. Learning
position weight matrices from sequence and expression data. Acceptted by CSB 2007.
|If you have any question or comment with this
program, please send me