Table 1.

Comparison of ranked target lists between the 50 b every 50-b and the 36 b every 36-b array platforms

Comparison of ranked target lists for the top 75 targets from the 50 b every 50-b array data set with the top 39 targets from the 36 b every 36-b array data set. (A) The upper panel displays the false-positive rates (FPR) calculated for each list considered separately. (B) The lower panel displays the results after merging the list of the top 75 targets from the 50-b arrays with the list of the top 39 targets from the 36-b arrays. The comparison is performed (see Methods for full details) by first creating the union of the two separate lists and then counting the number of union target regions specific to either the 50-b array data set or the 36-b array data set, or those targets identified by both platforms. In each of these three categories, the union regions that were tested for validation are displayed as well as the associated false-positive rate (FPR). Supplemental Table 2 is a similar comparison between the 50-b and 36-b array data with a more restrictive list of targets (top 25 targets) from the 36-b array (with a lower false-positive rate). The FPR is defined as TN/(TN + TP), where TN are the true negatives and TP are the true positives from the STAT1 ChIP-PCR analysis for target validations. Note that the FPR plus the PPV (positive predictive value as discussed in Fig. 2) sum to 1.