A new biomarker concept to predict human health effects of nutrients and develop nutraceuticals based on systems and network biology is presented here. We characterize the pathway responses of S. cerevisiae upon defined perturbations: controlled environmental stimuli using an antioxidant model compound, deletion of a gene whose Folinic acid calcium salt pentahydrate protein product constitutes a significant node of the network architecture and insertion of its human ortholog, and we assess their interaction by integrating such information into graphical network models, which elucidate predictive hypotheses to explain emergent behaviors. We further investigated the potential connectivity among these sub-networks. To do so, we extracted from the primary FA-specific network the first neighborhood of nodes, which appeared as significantly differentially expressed in the original gene list, and then we imported the corresponding small networks in Cytoscape and checked their connectivity to the larger network. This procedure was followed separately for each sub-network. In this manner, we observed that the larger sub-network was interconnected to only one of the remaining three sub-networks, while the resulting network was named LOUREIRIN-B ACTMOD network and is shown in Figure 2C. Subsequently, the ACTMOD network was studied for node sub-cellular localization. As indicated in Figure 2D, nodes were localized 
in almost every cellular compartment, however, most of the nodes belonged to the yeast mitochondrion, where the majority of reactive oxygen species species is generated. To understand the functional connectivity of the ACTMOD network, we determined the GO terms being significantly overrepresented. To be clinically relevant, results as those obtained above need to be translated and reduced to the level of a testable hypothesis about individual genes and proteins within the condition of interest. By integrating information from network connectivity and gene expression data, a list of 13 genes 2present in the MCODE clusters, active modules and the ANOVA list of significant genes2 was obtained. We examined the transcriptional regulation of these set of genes, and as shown in Figure 3B, a very tight regulatory network controls the expression of all 13 genes with more than 20 transcription factors being involved. A literature survey on these TFs revealed that a fraction of them has a crucial role in different stress responses in yeast. A second set of transcriptional regulators is involved in the cell cycle progression, while the function of one TF is connected with the activation of genes involved in ethanol consumption, a phenotypic deficiency observed during our batch cultivations. A third group of TFs was one of the main coordinators of the fine-tuning of the yeast response to FA. Rpn4 encodes a transcription activator that induces the proteasome genes, and recent studies have led to a model in which the proteasome homeostasis is regulated by a negative feedback circuit, a mechanism that exists also in higher eukaryotes, including humans. In one of these studies, it was demonstrated that downregulation of the proteasome genes, regulated by Rpn4, was able to reduce the active proteasome levels, a finding with potential clinical relevance in cancer cells. Aft1 is another TF that controls the activation of 40% of the up-regulated genes in response to neurotoxicants. Recent studies have elucidated the mechanisms of neural damage associated with these compounds and their linkage to the development of Parkinsonism symptoms. Prd1p and Pdr3p together confer resistance to several drugs through transcriptional activation of ABC transporter genes and members of the major facilitator super- family of drug efflux pumps, resulting in the expulsion of various structurally unrelated molecules. Pdr1p directly binds to xenobiotics to activate genes, and interacts physically and functionally with the Gal11p/MED15 subunit of the Mediator. The Mediator coactivator interacts with RNA polymerase II, which is in agreement with our observation.