Enhancer analysis is as much an art as a science. Multiple
sequence alignments attempt to represent an accurate summary of
the evolutionary
relationship between sequences, but ambiguities about positioning
inevitably lead to mis-aligned sequences. Automated alignments
make objective
judgments about the "optimal" placement of gaps, but additional
knowledge, such as binding sites for known regulators, can reveal
mistakes.
Similarly, choosing the right set of motifs for DNA-binding
proteins such as transcription factors, with the proper consensus
sequences and thresholds
for mismatches from the optimal sequence, is a trial-and-error
process. Twine is designed to allow experimentation of stringency
thresholds and the
set of motifs to consider, as well as the accuracy and
completeness of the alignment, by supporting dynamic adjustments
of threshold parameters.
Both consensus sequences and matrices are supported for motifs, so
data from multiple sources can be incorporated.
