Additional files to: Valledor, Luis, Takeshi Furuhashi, Luis Recuenco-Muñoz, Stefanie Wienkoop, and Wolfram Weckwerth. 2014. "System-Level Network Analysis of Nitrogen Starvation and Recovery in Chlamydomonas Reinhardtii Reveals Potential New Targets for Increased Lipid Accumulation." Biotechnology for Biofuels 7 (1): 171. doi:10.1186/s13068-014-0171-1. Additional file 1: Figure S1.: Comparison of cell growth (fresh weight) of N repleted and depleted cultures of C. reinhardtii CC503 grown under the same conditions employed in this study. Cultures were grown in triplicate. Additional file 2: Table S1.: List of the 1,534 quantified proteins in whole cell fractions. Protein abundance was determined following an NSAF approach. The mean abundance ± SD for each sampling time is indicated, as well as P- and q-values. Percentage of coverage and number of unique peptides used for identification and score are indicated. Deflines, E.C. number, and MapMan bins were manually curated. Table S2. List of proteins extracted from nuclei-enriched extracts. Spectra were processed as previously described. Deflines, E.C. number, and MapMan bins were manually curated. Table S3. List of the 52 quantified metabolites. Metabolite abundance was estimated from the peak areas of the indicated characteristic ions. The mean normalized abundances with respect to control ± SD for each sampling time are indicated as well as P-values. Table S4. PCA summary, sample scores, and variable loadings of the integrated (proteins, metabolites, physiological measurements) dataset. Table S5. PLS-discriminant analysis correlations of the integrated dataset. Table S6. sPLS analysis correlations of the integrated dataset employing proteins-metabolites as predictive variables and physiological measurements as responsive variables. Table S7. Values of individual replicates of the datasets employed for multivariate analyses. (a) Normalized abundances for proteins; (b, c) normalized abundances for polar metabolites and FAME; (d) physiological measurements. Table S8. List of nuclear proteins obtained after the joint analysis of total protein and nuclear extracts. Spectra were processed as previously described, but one step of in silico fractionation was added to remove non-nuclear contaminants. Deflines, E.C. number, and MapMan bins were manually curated. Table S9. Correspondence between 454 dataset (Miller et al.[33]), Chlamydomonas 5.3, and MapMan analyses. Abundace of 454 detected transcripts is also provided. Table S10. Granger causality analyzed by COVAIN (Sun and Weckwerth, 2012 [38]) between metabolites, proteins, and physiological data. Additional file 3: Figure S2.: Representation of nitrogen starvation- and recovery-induced changes in photosynthetic pathways using MapMan. Individual plots show the variations in protein abundance at the six studied time points. Only differential proteins (P < 0.05) were plotted. Protein abundances were normalized as a percentage of the maximal value in the time series. Additional file 4: Figure S3.: Representation of simplified N transport, fixation, and assimilation pathways showing N starvation- and recovery-induced changes using MapMan. Individual plots show the variations in protein abundance at the six studied time points. Only differential proteins (P < 0.05) were plotted. Protein abundances were normalized as a percentage of the maximal value in the time series. Additional file 5: Figure S4.: Representation of ribosomal proteins using MapMan. Individual plots show the variations in protein abundance at the six studied time points. Only differential proteins (P < 0.05) were plotted. Protein abundances were normalized as a percentage of the maximal value in the time series. Additional file 6: Figure S5.: Metabolic changes in lipid biosynthetic pathways induced by N starvation. (A) MapMan-based representation of the evolution of the enzymes related to fatty acid biosynthesis and elongation. (B) Variations in the protein abundance of the functional groups involved in lipid metabolism. (C) Evolution of the cellular content of the indicated fatty acids during the time course experiment. (D) sPLS-based network showing the most correlated proteins and metabolites to glycerol and C18:2 nodes. C18:2 was connected to ammonia transport and assimilation proteins, whereas glycerol was mainly connected to glycerol biosynthesis proteins and NADH:ubiquinone oxidoreductases. Interestingly, Cre09.g405500.t1.3 (major lipid droplet protein, with a high importance in the model) links C18:2 and glycerol. This protein was manually annotated to major lipid droplet protein. Other proteins that link both metabolites are a ribosomal protein and g13518.t1 (annotated as amine oxidase). Network interpretation: edge color indicates correlation between nodes, circle diameter indicates the importance of each node within the network, and node color its radiality. Figure was plotted employing Cytoscape 2.8.3. Additional file 7: Figure S6.: Representation of nitrogen starvation- and recovery-induced changes in nucleotide metabolism pathways. Individual plots show the variations in protein abundance at the six studied time points. Only differential proteins (P < 0.05) were plotted. Protein abundances were normalized as a percentage of the maximal value in the time series. Additional file 8: Figure S7.: Classification of the integrated dataset employing different multivariate methods such as principal component analysis, PCA (a,d), partial least squares discriminant analysis, PLS-DA (b,e), and sparse partial least squares, s-PLS (c,f), for which proteins and metebolites were used as predictive variables and physiological measurements as explanatory. For all analyses the first two components allowed an effective classification of the samples based on the sample time. Additional file 9: Figure S8.: Simplified sPLS-based network constructed considering only highly correlated nodes (correlation > |0.85|). The three main nodes established different correlations among them, with the glycerol-C18:2 cluster negatively correlated to the others. Interestingly, the fresh weight (FW) cluster is positively linked to the amino acids-myo-inositol cluster by a bridge constituted by the branched amino acids. Val, Leu, and Ile are positively correlated to a monofunctional catalase grouping with amino acids, and on the other side to an acyl-CoA oxidase linked to the FW. Surprisingly one enoyl-ACP-reductase, which is correlated to the amino acids cluster, is negatively correlated to glycerol-C18:2. Glycerol is also negatively correlated to Cre12.g536800.t1.2, a protein showing an uncharacterized domain which is related to FW. Network interpretation: edge color indicates correlation between nodes, circle diameter indicates the importance of each node within the network, and node color its radiality. Figure was plotted employing Cytoscape 2.8.3. Additional file 10: Figure S9.: Detailed representation of the fresh weight (FW). FW is correlated to different kinds of enzymes: signaling enzymes, mainly GTPases and Ca2+ dependent, endopeptidases, oxidoreductases, and acyl-CoA synthases, oxidases, and dehydrogenases. Node size is dependent on stress value and node color on radiality. Edge color indicates correlation (red, positive, green negative). Figures were plotted employing Cytoscape 2.8.3. Only highly correlated nodes (correlation > |0.85|) were plotted. Additional file 11: Figure S10.: Detailed representation of the amino acid cluster. This cluster showed complex interactions, in which nitrogen assimilation and amino acid metabolism enzymes interact with lipid biosynthesis proteins. Based on this network, some relation between N metabolism and lipid biosynthesis can be established, for example, at the level of one enoyl-ACP-reductase, which correlated to Asp, Arg, Glu, and myo-inositol, or a 3-oxoacyl-ACP reductase linked to Thr and myo-inositol. Thus, myo-inositol is a metabolite that needs to be further investigated since it is in the center of this cluster, linking amino acids, lipids, and sugars. Node size is dependent on stress value and node color on radiality. Edge color indicates correlation (red, positive, green negative). Figures were plotted employing Cytoscape 2.8.3. Only highly correlated nodes (correlation > |0.85|) were plotted.