Figure 1.

Expression change of pyoverdin genes. We analyze some expression data of pyoverdin genes (pvdD and pvdJ) of public database (GEO and Array- Express) by commonly used statistical methods (log-FC, RankProducts, t-rank and SAM). The threshold value of log-FC is set to 2 (4-fold) and that of RankProducts, t-rank and SAM are set to upper 300 gene. All dataset are normalized by RMA method separately. If both genes are co-expressed, corresponding box is filled in white, otherwise gray. Figure shows that pyoverdin genes are prone to be detected in any experiment condition and our method focuses on much experiment condition specific DEGs (GSE7704).

Tsuyuzaki et al. BMC Genomics 2013 14(Suppl 2):S9   doi:10.1186/1471-2164-14-S2-S9