Figure 3
Mining combinatorial features: identifying trans-factor clusters. The average size (i.e. the number of trans-factors members) of the identified clusters is displayed at different combinations of k and Ï„ values. (A) The white spot marks the configuration of the parameters selected to extract the clusters in the presence of recurrent trans-factors. (B) For each trans-factor, the number of occurrences in the identified clusters is plotted on the x axis, against the number of bound UTRs on the y axis. Recurrent trans-factors occurring in more than one cluster are labeled in orange. (C) After all the recurrent trans-factors were removed from the analysis, the average size of the identified clusters was displayed on the y axis at different combinations of k and Ï„ values. The white spot identifies the configuration of parameters selected to extract the clusters of sporadic trans-factors. (D) The proportions among the number of recurrent, sporadic, and absent trans-factors are shown. (E) The proportions among the number of interactions associated with recurrent, sporadic, and absent trans-factors are shown.