Identifying differentially activated pathways by augmenting activities of transcription factor target genes

摘要

There are many approaches to discovering significant pathways in expression profiling.Our objective is to develop a robust approach to capturing differentially expressed pathways in dis-ease expression profiling using pathways that have a transcription factor(TF). After collecting TFtarget gene, the TF target gene activity information in pathway and the pathway activity informationare calculated using a t test. The p values from two analyses are integrated by the Chi-square inversemethod. We evaluated our method using two cancer microarray data sets and comparing our resultswith those from two other pathway analysis methods. Our approach finds differentially expressedpathways that are directly associated with essential thrombocythemia, as well as general pathwaysthat reflect cancer type-nonspecific traits and cancer type-specific pathways in breast cancer data.To test the robustness of our approach, we analyzed two lung data sets. Our approach also pro-vides consistent and biologically reliable results. This study shows that it is better to find significantpathways in expression profiling where the TF target information is integrated within the pathwayactivity information. Our proposed approach is robust, and can also find cancer type-specific differ-entially expressed pathways.