Correlations of protein networks with physiological parameters
Correlation coefficient was determined between organismal measurements versus protein FC. Correlated variables included blood osmolality, condition factor (K), experimental salinity, and time at salinity above 65g/kg (~2X SW). Of these variables, blood osmolality difference from treatment to control was correlated with protein abundance with r2>0.9 for 446 proteins, with 280 of these proteins being significantly regulated in at least one treatment (Supplementary Table 2). Condition factor difference was correlated with r2>0.9 for 212 proteins, of which 125 were significantly regulated in at least one treatment. Experimental salinity was correlated with an r2>0.9 for only 16 proteins (6 significant) and time above 65g/kg was correlated with an r2>0.9 for 51 proteins (20 sig). Proteins with high blood osmolality correlation were analyzed using the STRING database, revealing one network with greater than three nodes which consisted of proteins involved in cell adhesion, ECM organization, and junctional complexes (Figure 5C). The STRING database annotated six of these proteins as associated with the serpin protein domain (SERine Protease Inhibitor), including serine protease inhibitor A3K, α-1-antitrypsin, protein Z-dependent protease inhibitor, angiotensinogen, complement factor D, and heparin cofactor 2. Four of the proteins were found in the “focal adhesion” KEGG molecular pathway, including α-actinin-1 X3, fibronectin, and two integrins (α-1 and α-6 X1). All proteins had negative slopes (decreasing FC with increasing blood osmolality) except for catenin δ-1 X3 and tropomyosin β X2, and the later was also found in a pejus range network.