< Back to Article

Chromatin Landscapes of Retroviral and Transposon Integration Profiles

Figure 4

Scale-based analysis of integration bias.

A) Association of the unselected integration profiles with various genome-wide features across different scales. Measure of association is a normalized t-score (see Material and Methods), computed on rank-normalized feature values, visualized on a blue-gray-red scale from negative to positive t-scores. Associations that are not significant (FDR-corrected ) are white. A positive (negative) t-score for a certain scale and feature means that for that particular feature, the mean values in a 200 bp window around the integrations are on average higher (lower) than the mean values in a 200 bp window around the points at a distance of upstream and downstream from the integration (see Material and Methods). The dendrogram shows a hierarchical clustering of the profiles using the euclidean distance measure and ward linkage. B) The rank-transformed smallest scale at which significance is achieved, with a scale going from white (small scale) to black (large scale). A feature is called a ‘macrofeature’ if its smallest significant scale is larger than the mean rank-normalized smallest significant scale across features, in both systems. C) Features associated with transcriptional repression and/or activation, based on published literature.

Figure 4