Using unsupervised hierarchical clustering with WGCNA consensus analysis we found twenty-four distinct modules (Fig. 1C). Note that two of the representative modules, light-yellow and turquoise, showed consistent gene expression correlations under three different conditions (during differentiation without EtOH treatment, EtOH treatment in undifferentiated hESCs and EtOH treatment in differentiated EBs) (Fig. 1C, indicated with the asterisks). Thus, the light-yellow module showed significant repression in differentiation, whereas a turquoise module exhibited the opposite, suggesting an essential role of genes in these modules during differentiation and also in EtOH-induced molecular alterations. The consensus module trait relationships are represented as correlation values and significance (Fig. S1). Furthermore, to delineate the top genes within these modules, we made a Volcano plot comparing significance and fold change (Fig. 1D). We analyzed natural clusters on the fourteen significant modules using the R package cluster Profiler (Fig. 1E) (Yu et al., 2012). It revealed that the modules formed distinct ontological clusters with little to no overlap. We also identified unique pathways for each module using the Kyoto Encyclopedia of Genes and Genomes (KEGG) Pathway (Fig. 1F) (Kanehisa
