Machine Learning Prediction of Adenovirus D8 Conjunctivitis Complications from Viral Whole-Genome Sequence.

Objective: To obtain complete DNA sequences of adenoviral (AdV) D8 genome from patients with conjunctivitis and determine the relation of sequence variation to clinical outcomes. Design: This study is a post hoc analysis of banked conjunctival swab samples from the BAYnovation Study, a previously conducted, randomized controlled clinical trial for AdV conjunctivitis. Participants: Ninety-six patients with AdV D8-positive conjunctivitis who received placebo treatment in the BAYnovation Study were included in the study. Methods: DNA from conjunctival swabs was purified and subjected to whole-genome viral DNA sequencing. Adenovirus D8 variants were identified and correlated with clinical outcomes, including 2 machine learning methods. Main Outcome Measures: Viral DNA sequence and development of subepithelial infiltrates (SEIs) were the main outcome measures. Results: From initial sequencing of 80 AdV D8-positive samples, full adenoviral genome reconstructions were obtained for 71. A total of 630 single-nucleotide variants were identified, including 156 missense mutations. Sequence clustering revealed 3 previously unappreciated viral clades within the AdV D8 type. The likelihood of SEI development differed significantly between clades, ranging from 83% for Clade 1 to 46% for Clade 3. Genome-wide analysis of viral single-nucleotide polymorphisms failed to identify single-gene determinants of outcome. Two machine learning models were independently trained to predict clinical outcome using polymorphic sequences. Both machine learning models correctly predicted development of SEI outcomes in a newly sequenced validation set of 16 cases (P = 1.5 × 10-5). Prediction was dependent on ensemble groups of polymorphisms across multiple genes. Conclusions: Adenovirus D8 has ≥ 3 prevalent molecular substrains, which differ in propensity to result in SEIs. Development of SEIs can be accurately predicted from knowledge of full viral sequence. These results suggest that development of SEIs in AdV D8 conjunctivitis is largely attributable to pathologic viral sequence variants within the D8 type and establishes machine learning paradigms as a powerful technique for understanding viral pathogenicity.

Tunable electroconductive decellularized extracellular matrix hydrogels for engineering human cardiac microphysiological systems.

Cardiomyocytes differentiated from human induced pluripotent stem cells (hiPSCs) offer tremendous potential when used to engineer human tissues for drug screening and disease modeling; however, phenotypic immaturity reduces assay reliability when translating in vitro results to clinical studies. To address this, we have developed hybrid hydrogels comprised of decellularized porcine myocardial extracellular matrix (dECM) and reduced graphene oxide (rGO) to provide a more instructive microenvironment for proper cell and tissue development. A tissue-specific protein profile was preserved post-decellularization, and through the modulation of rGO content and degree of reduction, the mechanical and electrical properties of the hydrogels could be tuned. Engineered heart tissues (EHTs) generated using dECM-rGO hydrogel scaffolds and hiPSC-derived cardiomyocytes exhibited significantly increased twitch forces and had increased expression of genes that regulate contractile function. Improvements in various aspects of electrophysiological function, such as calcium-handling, action potential duration, and conduction velocity, were also induced by the hybrid biomaterial. dECM-rGO hydrogels could also be used as a bioink to print cardiac tissues in a high-throughput manner, and these tissues were utilized to assess the proarrhythmic potential of cisapride. Action potential prolongation and beat interval irregularities was observed in dECM-rGO tissues at clinical doses of cisapride, indicating that the enhanced electrophysiological function of these tissues corresponded well with a capability to produce physiologically relevant drug responses.

TWIST1 Heterodimerization with E12 Requires Coordinated Protein Phosphorylation to Regulate Periostin Expression.

Diffuse invasion into adjacent brain matter by glioblastoma (GBM) is largely responsible for their dismal prognosis. Previously, we showed that the TWIST1 (TW) bHLH transcription factor and its regulated gene periostin (POSTN) promote invasive phenotypes of GBM cells. Since TW functional effects are regulated by phosphorylation and dimerization, we investigated how phosphorylation of serine 68 in TW regulates TW dimerization, POSTN expression, and invasion in glioma cells. Compared with wild-type TW, the hypophosphorylation mutant, TW(S68A), impaired TW heterodimerization with the E12 bHLH transcription factor and cell invasion in vitro but had no effect on TW homodimerization. Overexpression of TW:E12 forced dimerization constructs (FDCs) increased glioma cell invasion and upregulated pro-invasive proteins, including POSTN, in concert with cytoskeletal reorganization. By contrast, TW:TW homodimer FDCs inhibited POSTN expression and cell invasion in vitro. Further, phosphorylation of analogous PXSP phosphorylation sites in TW:E12 FDCs (TW S68 and E12 S139) coordinately regulated POSTN and PDGFRa mRNA expression. These results suggested that TW regulates pro-invasive phenotypes in part through coordinated phosphorylation events in TW and E12 that promote heterodimer formation and regulate downstream targets. This new mechanistic understanding provides potential therapeutic strategies to inhibit TW-POSTN signaling in GBM and other cancers.

Synthesizing Signaling Pathways from Temporal Phosphoproteomic Data.

We present a method for automatically discovering signaling pathways from time-resolved phosphoproteomic data. The Temporal Pathway Synthesizer (TPS) algorithm uses constraint-solving techniques first developed in the context of formal verification to explore paths in an interaction network. It systematically eliminates all candidate structures for a signaling pathway where a protein is activated or inactivated before its upstream regulators. The algorithm can model more than one hundred thousand dynamic phosphosites and can discover pathway members that are not differentially phosphorylated. By analyzing temporal data, TPS defines signaling cascades without needing to experimentally perturb individual proteins. It recovers known pathways and proposes pathway connections when applied to the human epidermal growth factor and yeast osmotic stress responses. Independent kinase mutant studies validate predicted substrates in the TPS osmotic stress pathway.

Development of Selected Reaction Monitoring Methods to Systematically Quantify Kinase Abundance and Phosphorylation Stoichiometry in Human Samples.

Protein phosphorylation, one of the most common types of post-translational modifications, is the central regulatory mechanism of cellular signaling networks. In human cells, thousands of proteins are continuously and dynamically phosphorylated and dephosphorylated at specific sites and times in response to external and internal stimuli. Reversible phosphorylation is facilitated by the action of two protein superfamilies: kinases and phosphatases. Kinases play an essential role in almost every relevant physiological process in human cells and their deregulation is linked to pathologies ranging from cancer to autoimmune diseases.Systematic identification of kinases expressed in a particular cell type, quantification of their abundance, and precise determination of their phosphorylation stoichiometry are essential to understand the cellular signaling networks and physiology of a sample. Our protocol outlines the steps to build and use a high-throughput, comprehensive, modular, and robust selected reaction monitoring (SRM) proteomics framework to facilitate quantification of the kinome state in research or clinical human samples.

Integrated systems biology analysis of KSHV latent infection reveals viral induction and reliance on peroxisome mediated lipid metabolism.

Kaposi's Sarcoma associated Herpesvirus (KSHV), an oncogenic, human gamma-herpesvirus, is the etiological agent of Kaposi's Sarcoma the most common tumor of AIDS patients world-wide. KSHV is predominantly latent in the main KS tumor cell, the spindle cell, a cell of endothelial origin. KSHV modulates numerous host cell-signaling pathways to activate endothelial cells including major metabolic pathways involved in lipid metabolism. To identify the underlying cellular mechanisms of KSHV alteration of host signaling and endothelial cell activation, we identified changes in the host proteome, phosphoproteome and transcriptome landscape following KSHV infection of endothelial cells. A Steiner forest algorithm was used to integrate the global data sets and, together with transcriptome based predicted transcription factor activity, cellular networks altered by latent KSHV were predicted. Several interesting pathways were identified, including peroxisome biogenesis. To validate the predictions, we showed that KSHV latent infection increases the number of peroxisomes per cell. Additionally, proteins involved in peroxisomal lipid metabolism of very long chain fatty acids, including ABCD3 and ACOX1, are required for the survival of latently infected cells. In summary, novel cellular pathways altered during herpesvirus latency that could not be predicted by a single systems biology platform, were identified by integrated proteomics and transcriptomics data analysis and when correlated with our metabolomics data revealed that peroxisome lipid metabolism is essential for KSHV latent infection of endothelial cells.

A computationally engineered RAS rheostat reveals RAS-ERK signaling dynamics.

Synthetic protein switches controlled with user-defined inputs are powerful tools for studying and controlling dynamic cellular processes. To date, these approaches have relied primarily on intermolecular regulation. Here we report a computationally guided framework for engineering intramolecular regulation of protein function. We utilize this framework to develop chemically inducible activator of RAS (CIAR), a single-component RAS rheostat that directly activates endogenous RAS in response to a small molecule. Using CIAR, we show that direct RAS activation elicits markedly different RAS-ERK signaling dynamics from growth factor stimulation, and that these dynamics differ among cell types. We also found that the clinically approved RAF inhibitor vemurafenib potently primes cells to respond to direct wild-type RAS activation. These results demonstrate the utility of CIAR for quantitatively interrogating RAS signaling. Finally, we demonstrate the general utility of our approach in design of intramolecularly regulated protein tools by applying it to the Rho family of guanine nucleotide exchange factors.

Endogenous N-terminal Domain Cleavage Modulates α1D-Adrenergic Receptor Pharmacodynamics.

The α1D-adrenergic receptor (ADRA1D) is a key regulator of cardiovascular, prostate, and central nervous system functions. This clinically relevant G protein-coupled receptor has proven difficult to study, as it must form an obligate modular homodimer containing the PDZ proteins scribble and syntrophin or become retained in the endoplasmic reticulum as non-functional protein. We previously determined that targeted removal of the N-terminal (NT) 79 amino acids facilitates ADRA1D plasma membrane expression and agonist-stimulated functional responses. However, whether such an event occurs in physiological contexts was unknown. Herein, we report the ADRA1D is subjected to innate NT processing in cultured human cells. SNAP near-infrared imaging and tandem-affinity purification revealed the ADRA1D is expressed as both full-length and NT truncated forms in multiple human cell lines. Serial truncation mapping identified the cleavage site as Leu(90)/Val(91) in the 95-amino acid ADRA1D NT domain, suggesting human cells express a Δ1-91 ADRA1D species. Tandem-affinity purification MS/MS and co-immunoprecipitation analysis indicate NT processing of ADRA1D is not required to form scribble-syntrophin macromolecular complexes. Yet, label-free dynamic mass redistribution signaling assays demonstrate that Δ1-91 ADRA1D agonist responses were greater than WT ADRA1D. Mutagenesis of the cleavage site nullified the processing event, resulting in ADRA1D agonist responses less than the WT receptor. Thus, we propose that processing of the ADRA1D NT domain is a physiological mechanism employed by cells to generate a functional ADRA1D isoform with optimal pharmacodynamic properties.

Methods for the Analysis of Protein Phosphorylation-Mediated Cellular Signaling Networks.

Protein phosphorylation-mediated cellular signaling networks regulate almost all aspects of cell biology, including the responses to cellular stimulation and environmental alterations. These networks are highly complex and comprise hundreds of proteins and potentially thousands of phosphorylation sites. Multiple analytical methods have been developed over the past several decades to identify proteins and protein phosphorylation sites regulating cellular signaling, and to quantify the dynamic response of these sites to different cellular stimulation. Here we provide an overview of these methods, including the fundamental principles governing each method, their relative strengths and weaknesses, and some examples of how each method has been applied to the analysis of complex signaling networks. When applied correctly, each of these techniques can provide insight into the topology, dynamics, and regulation of protein phosphorylation signaling networks.

Technical advances in proteomics: new developments in data-independent acquisition.

The ultimate aim of proteomics is to fully identify and quantify the entire complement of proteins and post-translational modifications in biological samples of interest. For the last 15 years, liquid chromatography-tandem mass spectrometry (LC-MS/MS) in data-dependent acquisition (DDA) mode has been the standard for proteomics when sampling breadth and discovery were the main objectives; multiple reaction monitoring (MRM) LC-MS/MS has been the standard for targeted proteomics when precise quantification, reproducibility, and validation were the main objectives. Recently, improvements in mass spectrometer design and bioinformatics algorithms have resulted in the rediscovery and development of another sampling method: data-independent acquisition (DIA). DIA comprehensively and repeatedly samples every peptide in a protein digest, producing a complex set of mass spectra that is difficult to interpret without external spectral libraries. Currently, DIA approaches the identification breadth of DDA while achieving the reproducible quantification characteristic of MRM or its newest version, parallel reaction monitoring (PRM). In comparative de novo identification and quantification studies in human cell lysates, DIA identified up to 89% of the proteins detected in a comparable DDA experiment while providing reproducible quantification of over 85% of them. DIA analysis aided by spectral libraries derived from prior DIA experiments or auxiliary DDA data produces identification and quantification as reproducible and precise as that achieved by MRM/PRM, except on low­abundance peptides that are obscured by stronger signals. DIA is still a work in progress toward the goal of sensitive, reproducible, and precise quantification without external spectral libraries. New software tools applied to DIA analysis have to deal with deconvolution of complex spectra as well as proper filtering of false positives and false negatives. However, the future outlook is positive, and various researchers are working on novel bioinformatics techniques to address these issues and increase the reproducibility, fidelity, and identification breadth of DIA.

Dynamic mass redistribution reveals diverging importance of PDZ-ligands for G protein-coupled receptor pharmacodynamics.

G protein-coupled receptors (GPCRs) are essential membrane proteins that facilitate cell-to-cell communication and co-ordinate physiological processes. At least 30 human GPCRs contain a Type I PSD-95/DLG/Zo-1 (PDZ) ligand in their distal C-terminal domain; this four amino acid motif of X-[S/T]-X-[φ] sequence facilitates interactions with PDZ domain-containing proteins. Because PDZ protein interactions have profound effects on GPCR ligand pharmacology, cellular localization, signal-transduction effector coupling and duration of activity, we analyzed the importance of Type I PDZ ligands for the function of 23 full-length and PDZ-ligand truncated (ΔPDZ) human GPCRs in cultured human cells. SNAP-epitope tag polyacrylamide gel electrophoresis revealed most Type I PDZ GPCRs exist as both monomers and multimers; removal of the PDZ ligand played minimal role in multimer formation. Additionally, SNAP-cell surface staining indicated removal of the PDZ ligand had minimal effects on plasma membrane localization for most GPCRs examined. Label-free dynamic mass redistribution functional responses, however, revealed diverging effects of the PDZ ligand. While no clear trend was observed across all GPCRs tested or even within receptor families, a subset of GPCRs displayed diminished agonist efficacy in the absence of a PDZ ligand (i.e. HT2RB, ADRB1), whereas others demonstrated enhanced agonist efficacies (i.e. LPAR2, SSTR5). These results demonstrate the utility of label-free functional assays to tease apart the contributions of conserved protein interaction domains for GPCR signal-transduction coupling in cultured cells.

Individual protomers of a G protein-coupled receptor dimer integrate distinct functional modules.

Recent advances in proteomic technology reveal G-protein-coupled receptors (GPCRs) are organized as large, macromolecular protein complexes in cell membranes, adding a new layer of intricacy to GPCR signaling. We previously reported the α1D-adrenergic receptor (ADRA1D)-a key regulator of cardiovascular, urinary and CNS function-binds the syntrophin family of PDZ domain proteins (SNTA, SNTB1, and SNTB2) through a C-terminal PDZ ligand interaction, ensuring receptor plasma membrane localization and G-protein coupling. To assess the uniqueness of this novel GPCR complex, 23 human GPCRs containing Type I PDZ ligands were subjected to TAP/MS proteomic analysis. Syntrophins did not interact with any other GPCRs. Unexpectedly, a second PDZ domain protein, scribble (SCRIB), was detected in ADRA1D complexes. Biochemical, proteomic, and dynamic mass redistribution analyses indicate syntrophins and SCRIB compete for the PDZ ligand, simultaneously exist within an ADRA1D multimer, and impart divergent pharmacological properties to the complex. Our results reveal an unprecedented modular dimeric architecture for the ADRA1D in the cell membrane, providing unexpected opportunities for fine-tuning receptor function through novel protein interactions in vivo, and for intervening in signal transduction with small molecules that can stabilize or disrupt unique GPCR:PDZ protein interfaces.

The Tec Kinase-Regulated Phosphoproteome Reveals a Mechanism for the Regulation of Inhibitory Signals in Murine Macrophages.

Previous work has shown conflicting roles for Tec family kinases in regulation of TLR-dependent signaling in myeloid cells. In the present study, we performed a detailed investigation of the role of the Tec kinases Btk and Tec kinases in regulating TLR signaling in several types of primary murine macrophages. We demonstrate that primary resident peritoneal macrophages deficient for Btk and Tec secrete less proinflammatory cytokines in response to TLR stimulation than do wild-type cells. In contrast, we found that bone marrow-derived and thioglycollate-elicited peritoneal macrophages deficient for Btk and Tec secrete more proinflammatory cytokines than do wild-type cells. We then compared the phosphoproteome regulated by Tec kinases and LPS in primary peritoneal and bone marrow-derived macrophages. From this analysis we determined that Tec kinases regulate different signaling programs in these cell types. In additional studies using bone marrow-derived macrophages, we found that Tec and Btk promote phosphorylation events necessary for immunoreceptor-mediated inhibition of TLR signaling. Taken together, our results are consistent with a model where Tec kinases (Btk, Tec, Bmx) are required for TLR-dependent signaling in many types of myeloid cells. However, our data also support a cell type-specific TLR inhibitory role for Btk and Tec that is mediated by immunoreceptor activation and signaling via PI3K.

Protein kinase PKN1 represses Wnt/ß-catenin signaling in human melanoma cells.

Advances in phosphoproteomics have made it possible to monitor changes in protein phosphorylation that occur at different steps in signal transduction and have aided the identification of new pathway components. In the present study, we applied this technology to advance our understanding of the responses of melanoma cells to signaling initiated by the secreted ligand WNT3A. We started by comparing the phosphopeptide patterns of cells treated with WNT3A for different periods of time. Next, we integrated these data sets with the results from a siRNA screen that targeted protein kinases. This integration of siRNA screening and proteomics enabled us to identify four kinases that exhibit altered phosphorylation in response to WNT3A and that regulate a luciferase reporter of ß-catenin-responsive transcription (ß-catenin-activated reporter). We focused on one of these kinases, an atypical PKC kinase, protein kinase N1 (PKN1). Reducing the levels of PKN1 with siRNAs significantly enhances activation of ß-catenin-activated reporter and increases apoptosis in melanoma cell lines. Using affinity purification followed by mass spectrometry, we then found that PKN1 is present in a protein complex with a WNT3A receptor, Frizzled 7, as well as with proteins that co-purify with Frizzled 7. These data establish that the protein kinase PKN1 inhibits Wnt/ß-catenin signaling and sensitizes melanoma cells to cell death stimulated by WNT3A.

Receptor tyrosine kinases fall into distinct classes based on their inferred signaling networks.

Although many anticancer drugs that target receptor tyrosine kinases (RTKs) provide clinical benefit, their long-term use is limited by resistance that is often attributed to increased abundance or activation of another RTK that compensates for the inhibited receptor. To uncover common and unique features in the signaling networks of RTKs, we measured time-dependent signaling in six isogenic cell lines, each expressing a different RTK as downstream proteins were systematically perturbed by RNA interference. Network models inferred from the data revealed a conserved set of signaling pathways and RTK-specific features that grouped the RTKs into three distinct classes: (i) an EGFR/FGFR1/c-Met class constituting epidermal growth factor receptor, fibroblast growth factor receptor 1, and the hepatocyte growth factor receptor c-Met; (ii) an IGF-1R/NTRK2 class constituting insulin-like growth factor 1 receptor and neurotrophic tyrosine receptor kinase 2; and (iii) a PDGFRß class constituting platelet-derived growth factor receptor ß. Analysis of cancer cell line data showed that many RTKs of the same class were coexpressed and that increased abundance of an RTK or its cognate ligand frequently correlated with resistance to a drug targeting another RTK of the same class. In contrast, abundance of an RTK or ligand of one class generally did not affect sensitivity to a drug targeting an RTK of a different class. Thus, classifying RTKs by their inferred networks and then therapeutically targeting multiple receptors within a class may delay or prevent the onset of resistance.

A general molecular affinity strategy for global detection and proteomic analysis of lysine methylation.

Lysine methylation of histone proteins regulates chromatin dynamics and plays important roles in diverse physiological and pathological processes. However, beyond histone proteins, the proteome-wide extent of lysine methylation remains largely unknown. We have engineered the naturally occurring MBT domain repeats of L3MBTL1 to serve as a universal affinity reagent for detecting, enriching, and identifying proteins carrying a mono- or dimethylated lysine. The domain is broadly specific for methylated lysine ("pan-specific") and can be applied to any biological system. We have used our approach to demonstrate that SIRT1 is a substrate of the methyltransferase G9a both in vitro and in cells, to perform proteome-wide detection and enrichment of methylated proteins, and to identify candidate in-cell substrates of G9a and the related methyltransferase GLP. Together, our results demonstrate a powerful new approach for global and quantitative analysis of methylated lysine, and they represent the first systems biology understanding of lysine methylation.

A disease-associated PTPN22 variant promotes systemic autoimmunity in murine models.

Multiple autoimmune diseases, including type 1 diabetes, rheumatoid arthritis, Graves disease, and systemic lupus erythematosus, are associated with an allelic variant of protein tyrosine phosphatase nonreceptor 22 (PTPN22), which encodes the protein LYP. To model the human disease-linked variant LYP-R620W, we generated knockin mice expressing the analogous mutation, R619W, in the murine ortholog PEST domain phosphatase (PEP). In contrast with a previous report, we found that this variant exhibits normal protein stability, but significantly alters lymphocyte function. Aged knockin mice exhibited effector T cell expansion and transitional, germinal center, and age-related B cell expansion as well as the development of autoantibodies and systemic autoimmunity. Further, PEP-R619W affected B cell selection and B lineage-restricted variant expression and was sufficient to promote autoimmunity. Consistent with these features, PEP-R619W lymphocytes were hyperresponsive to antigen-receptor engagement with a distinct profile of tyrosine-phosphorylated substrates. Thus, PEP-R619W uniquely modulates T and B cell homeostasis, leading to a loss in tolerance and autoimmunity.

Quantitative proteomic analysis of HIV-1 infected CD4+ T cells reveals an early host response in important biological pathways: protein synthesis, cell proliferation, and T-cell activation.

Human immunodeficiency virus (HIV-1) depends upon host-encoded proteins to facilitate its replication while at the same time inhibiting critical components of innate and/or intrinsic immune response pathways. To characterize the host cell response on protein levels in CD4+ lymphoblastoid SUP-T1 cells after infection with HIV-1 strain LAI, we used mass spectrometry (MS)-based global quantitation with iTRAQ (isobaric tag for relative and absolute quantification). We found 266, 60 and 22 proteins differentially expressed (DE) (P-value ≤ 0.05) at 4, 8, and 20 hours post-infection (hpi), respectively, compared to time-matched mock-infected samples. The majority of changes in protein abundance occurred at an early stage of infection well before the de novo production of viral proteins. Functional analyses of these DE proteins showed enrichment in several biological pathways including protein synthesis, cell proliferation, and T-cell activation. Importantly, these early changes before the time of robust viral production have not been described before.

Lysate microarrays enable high-throughput, quantitative investigations of cellular signaling.

Lysate microarrays (reverse-phase protein arrays) hold great promise as a tool for systems-level investigations of signaling and multiplexed analyses of disease biomarkers. To date, however, widespread use of this technology has been limited by questions concerning data quality and the specificity of detection reagents. To address these concerns, we developed a strategy to identify high-quality reagents for use with lysate microarrays. In total, we tested 383 antibodies for their ability to quantify changes in protein abundance or modification in 20 biological contexts across 17 cell lines. Antibodies yielding significant differences in signal were further evaluated by immunoblotting and 82 passed our rigorous criteria. The large-scale data set from our screen revealed that cell fate decisions are encoded not just by the identities of proteins that are activated, but by differences in their signaling dynamics as well. Overall, our list of validated antibodies and associated protocols establish lysate microarrays as a robust tool for systems biology.

Maximum entropy reconstructions of dynamic signaling networks from quantitative proteomics data.

Advances in mass spectrometry among other technologies have allowed for quantitative, reproducible, proteome-wide measurements of levels of phosphorylation as signals propagate through complex networks in response to external stimuli under different conditions. However, computational approaches to infer elements of the signaling network strictly from the quantitative aspects of proteomics data are not well established. We considered a method using the principle of maximum entropy to infer a network of interacting phosphotyrosine sites from pairwise correlations in a mass spectrometry data set and derive a phosphorylation-dependent interaction network solely from quantitative proteomics data. We first investigated the applicability of this approach by using a simulation of a model biochemical signaling network whose dynamics are governed by a large set of coupled differential equations. We found that in a simulated signaling system, the method detects interactions with significant accuracy. We then analyzed a growth factor mediated signaling network in a human mammary epithelial cell line that we inferred from mass spectrometry data and observe a biologically interpretable, small-world structure of signaling nodes, as well as a catalog of predictions regarding the interactions among previously uncharacterized phosphotyrosine sites. For example, the calculation places a recently identified tumor suppressor pathway through ARHGEF7 and Scribble, in the context of growth factor signaling. Our findings suggest that maximum entropy derived network models are an important tool for interpreting quantitative proteomics data.