Fig. 1.

The Gene Sort criteria interface for mining the Gene Expression Data (a). Options are available to filter probe results to prioritise a disease stage using the “Points per % change” and a disease “Direction of Progression” as expression increasing, decreasing or either. Results can be further filtered for consistency by applying the “Patient inconsistency penalty”. The Gene Expression Data interface displaying results at the probe level (b). The ranked gene transcript probes are presented on the left as a scrollable list and results for the selected probe ADXECCRS.32699_at are displayed. The “Apply Filter” feature can be used to search the ranked list of probes. Expression values are provided for each patient at each disease stage as a data matrix. Expression trends in relation to Progression across disease stages are displayed for patients as line graphs and boxplots. The boxplot represents the median and interquartile range with the maximum and minimum values shown. The percentage change in expression between disease stages NFT to STIC and STIC to HGSC are also displayed. Above the boxplots panel, the ranking score for the selected probe is displayed as 376.798.