Description

ENCODE Regulatory Elements from Boston University. This track annotates regulatory elements in all ENCODE regions, by combining evidence from multiple sources. This track can be used to look for novel promoters.

Methods

The lab used Naive Bayes to combine multiple experiments to produce a list of regulatory elements.

The evidence is listed below. For each input set, the Weng lab determine a cutoff score at the point of maximum correlation with a well-defined set of positive and negative examples of promoter regions. They combined the binarized datasets at the base-pair level using a Naive Bayes setting to produce a combined score which is binarized at p less than 0.01. Contiguous runs of predicted positions are clustered.

Credits

This work was done in collaboration at the Zhiping Weng lab at Boston University.