Protein-protein Interaction Data Research Articles

SLiM-Enrich: computational assessment of protein-protein interaction data as a source of domain-motif interactions.

Many important cellular processes involve protein–protein interactions (PPIs) mediated by a Short Linear Motif (SLiM) in one protein interacting with a globular domain in another. Despite their significance, these domain-motif interactions (DMIs) are typically low affinity, which makes them challenging to identify by classical experimental approaches, such as affinity pulldown mass spectrometry (AP-MS) and yeast two-hybrid (Y2H). DMIs are generally underrepresented in PPI networks as a result. A number of computational methods now exist to predict SLiMs and/or DMIs from experimental interaction data but it is yet to be established how effective different PPI detection methods are for capturing these low affinity SLiM-mediated interactions. Here, we introduce a new computational pipeline (SLiM-Enrich) to assess how well a given source of PPI data captures DMIs and thus, by inference, how useful that data should be for SLiM discovery. SLiM-Enrich interrogates a PPI network for pairs of interacting proteins in which the first protein is known or predicted to interact with the second protein via a DMI. Permutation tests compare the number of known/predicted DMIs to the expected distribution if the two sets of proteins are randomly associated. This provides an estimate of DMI enrichment within the data and the false positive rate for individual DMIs. As a case study, we detect significant DMI enrichment in a high-throughput Y2H human PPI study. SLiM-Enrich analysis supports Y2H data as a source of DMIs and highlights the high false positive rates associated with naïve DMI prediction. SLiM-Enrich is available as an R Shiny app. The code is open source and available via a GNU GPL v3 license at: https://github.com/slimsuite/SLiMEnrich. A web server is available at: http://shiny.slimsuite.unsw.edu.au/SLiMEnrich/.

Hippocampal transcriptome profiling combined with protein-protein interaction analysis elucidates Alzheimer's disease pathways and genes

Knowledge about the molecular mechanisms driving Alzheimer's disease (AD) is still limited. To learn more about AD biology, we performed whole transcriptome sequencing on the hippocampus of 20 AD cases and 10 age- and sex-matched cognitively healthy controls. We observed 2716 differentially expressed genes, of which 48% replicated in a second data set of 84 AD cases and 33 controls. We used an integrative network-based approach for combining transcriptomic and protein-protein interaction data to find differentially expressed gene modules that may reflect key processes in AD biology. A total of 735 differentially expressed genes were clustered into 33 modules, of which 82% replicated in a second data set, highlighting the robustness of this approach. These 27 modules were enriched for signal transduction, transport, response to stimulus, and several organic and cellular metabolic pathways. Ten modules interacted with previously described AD genes. Our study indicates that analyzing RNA-expression data based on annotated gene modules is more robust than on individual genes. We provide a comprehensive overview of the biological processes involved in AD, and the detected differentially expressed gene modules may provide a molecular basis for future research into mechanisms underlying AD.

Exploring synergy in ayurveda and traditional Indian systems of medicine

Abstract Synergy is a phenomenon of life. It has been observed that combinatory treatments may achieve more stable “network responses’’ than single drugs. Drug combination treatments are routinely applied in an increasing number of disciplines in medicine. Single drug treatments are less effective than drug combinations in complex treatment regimes. Plant derived herbal drugs or phytopharmaceuticals can be considered a natural “model” for combinatory treatments. Specific compositions of multi-component mixtures represent combined actions of single substances. Ayurveda, the traditional medicine of India has played an important role in disease management and health for many centuries, their present frequent use is challenged by the necessity to determine their complex composition and their multitarget mode of action. Ancient Ayurvedic literature like Sarangadhar Samhita focuses on the idea of synergy or ‘samyoga’. Network pharmacology research focuses on excellent systems biology techniques like protein interaction, genomic expression and mRNA expression data, which are employed to gain insight into the mechanism of action (MoA) prediction and validation. The following brief review provides valuable perspective regarding the bridge of Ayurveda and synergy research using network pharmacology as the tool to understand the molecular combination mechanism.

Disturbed ribosome-related modules were associated with Kawasaki disease.

Kawasaki disease (KD) is an acute vasculitis in young children, with ambiguous etiology. Early diagnosis will decrease the risk of coronary artery aneurysms and contributes to the favourable prognosis. This work aimed to identify disease modules that accurately predict clinical outcome using a systemic module tracking method. Based on the transcriptional data and protein-protein interaction (PPI) data, we constructed the differential co-expression network for KD. Then, a systemic module tracking method was performed to extract KD-related modules that accurately predict clinical outcome from the differential co-expression network, according to two steps: key genes identification and module inference by key gene expansion. 16 key genes were identified based on their importance in differential co-expression network and most of them were ribosomal protein-related genes. With each key gene as initial gene, we identified 10 disease modules with high predictive accuracy. Function analysis found that these disease modules were related to one common pathway-ribosome pathway. Our study for the first time indicated that disturbed ribosome-related disease modules might contribute to the development of KD. These modules could be considered as novel contributors to the progression of KD, and potential diagnostic biomarkers for predicting the clinical outcome.

A Hybrid ACO-Graph Entropy for Functional Modules Detection from Protein–Protein Interaction Network

Recent high-throughput experiments have generated protein-protein interaction data on a genomic scale, yielding the complete protein-protein interaction network for several organisms. Various graph clustering algorithms have been applied to protein interaction networks for detecting protein functional modules. Although the previous algorithms are scalable and robust, their accuracy is still limited because of the complex connectivity found in protein interaction networks. The Ant Colony Optimization (ACO) Algorithm has been adapted for the protein functional module detection by modeling the problem as an optimization problem. The adapted ACO (ACO-PFMDA) has obtained feasible solution but not as magnificent as those reported in the literature. Some shortcomings were identified and addressed by proposing a Modified Ant Colony Optimization Algorithm (ACO-PFMDM), which introduces two new scheme for controlling the two main parameters of ACO to solve PFMDP. Experiments on one popular benchmark dataset namely Saccharomyces cerevisiae which taken from two popular databases DIP and MIPS has been performed. The experimental result have proved that ACO-PFMDM have improved the overall performance of protein functional module detection. The search process of ACO-PFMDM has converged effectively compared to some state-of-art algorithms. Moreover, the proposed dynamic update of the heuristic parameters based on entropy has generated high quality tours and it can guide ants toward the effective solutions space in the initial search stages.

Fold change based approach for identification of significant network markers in breast, lung and prostate cancer.

Cancer belongs to a class of highly aggressive diseases and a leading cause of death in the world. With more than 100 types of cancers, breast, lung and prostate cancer remain to be the most common types. To identify essential network markers (NMs) and therapeutic targets in these cancers, the authors present a novel approach which uses gene expression data from microarray and RNA‐seq platforms and utilises the results from this data to evaluate protein–protein interaction (PPI) network. Differentially expressed genes (DEGs) are extracted from microarray data using three different statistical methods in R, to produce a consistent set of genes. Also, DEGs are extracted from RNA‐seq data for the same three cancer types. DEG sets found to be common in both platforms are obtained at three fold change (FC) cut‐off levels to accurately identify the level of change in expression of these genes in all three cancers. A cancer network is built using PPI data characterising gene sets at log‐FC (LFC)>1, LFC>1.5 and LFC>2, and interconnection between principal hub nodes of these networks is observed. Resulting network of hubs at three FC levels highlights prime NMs with high confidence in multiple cancers as validated by Gene Ontology functional enrichment and maximal complete subgraphs from CFinder.

Protein interaction and functional data indicate MTHFD2 involvement in RNA processing and translation

BackgroundThe folate-coupled metabolic enzyme MTHFD2 is overexpressed in many tumor types and required for cancer cell proliferation, and is therefore of interest as a potential cancer therapeutic target. However, recent evidence suggests that MTHFD2 has a non-enzymatic function which may underlie the dependence of cancer cells on this protein. Understanding this non-enzymatic function is important for optimal targeting of MTHFD2 in cancer.MethodsTo identify potential non-enzymatic functions of MTHFD2, we defined its interacting proteins using co-immunoprecipitation and mass spectrometry and integrated this information with large-scale co-expression analysis, protein dynamics, and gene expression response to MTHFD2 knockdown.ResultsWe found that MTHFD2 physically interacts with a set of nuclear proteins involved in RNA metabolism and translation, including components of the small ribosomal subunit and multiple members of the RNA-processing hnRNP family. Interacting proteins were also in general co-expressed with MTHFD2 in experiments that stimulate or repress proliferation, suggesting a close functional relationship. Also, unlike other folate one-carbon enzymes, the MTHFD2 protein has a short half-life and responds rapidly to serum. Finally, shRNA against MTHFD2 depletes several of its interactors and yields an overall transcriptional response similar to targeted inhibition of certain ribosomal subunits.ConclusionsTaken together, our findings suggest a novel function of MTHFD2 in RNA metabolism and translation.

Integration of Molecular Interactome and Targeted Interaction Analysis to Identify a COPD Disease Network Module

The polygenic nature of complex diseases offers potential opportunities to utilize network-based approaches that leverage the comprehensive set of protein-protein interactions (the human interactome) to identify new genes of interest and relevant biological pathways. However, the incompleteness of the current human interactome prevents it from reaching its full potential to extract network-based knowledge from gene discovery efforts, such as genome-wide association studies, for complex diseases like chronic obstructive pulmonary disease (COPD). Here, we provide a framework that integrates the existing human interactome information with experimental protein-protein interaction data for FAM13A, one of the most highly associated genetic loci to COPD, to find a more comprehensive disease network module. We identified an initial disease network neighborhood by applying a random-walk method. Next, we developed a network-based closeness approach (CAB) that revealed 9 out of 96 FAM13A interacting partners identified by affinity purification assays were significantly close to the initial network neighborhood. Moreover, compared to a similar method (local radiality), the CAB approach predicts low-degree genes as potential candidates. The candidates identified by the network-based closeness approach were combined with the initial network neighborhood to build a comprehensive disease network module (163 genes) that was enriched with genes differentially expressed between controls and COPD subjects in alveolar macrophages, lung tissue, sputum, blood, and bronchial brushing datasets. Overall, we demonstrate an approach to find disease-related network components using new laboratory data to overcome incompleteness of the current interactome.

A pipeline to translate glycosaminoglycan sequences into 3D models. Application to the exploration of glycosaminoglycan conformational space.

Mammalian glycosaminoglycans are linear complex polysaccharides comprising heparan sulfate, heparin, dermatan sulfate, chondroitin sulfate, keratan sulfate and hyaluronic acid. They bind to numerous proteins and these interactions mediate their biological activities. GAG-protein interaction data reported in the literature are curated mostly in MatrixDB database (http://matrixdb.univ-lyon1.fr/). However, a standard nomenclature and a machine-readable format of GAGs together with bioinformatics tools for mining these interaction data are lacking. We report here the building of an automated pipeline to (i) standardize the format of GAG sequences interacting with proteins manually curated from the literature, (ii) translate them into the machine-readable GlycoCT format and into SNFG (Symbol Nomenclature For Glycan) images and (iii) convert their sequences into a format processed by a builder generating three-dimensional structures of polysaccharides based on a repertoire of conformations experimentally validated by data extracted from crystallized GAG-protein complexes. We have developed for this purpose a converter (the CT23D converter) to automatically translate the GlycoCT code of a GAG sequence into the input file required to construct a three-dimensional model.

Predicting protein-protein interactions through sequence-based deep learning.

High-throughput experimental techniques have produced a large amount of protein-protein interaction (PPI) data, but their coverage is still low and the PPI data is also very noisy. Computational prediction of PPIs can be used to discover new PPIs and identify errors in the experimental PPI data. We present a novel deep learning framework, DPPI, to model and predict PPIs from sequence information alone. Our model efficiently applies a deep, Siamese-like convolutional neural network combined with random projection and data augmentation to predict PPIs, leveraging existing high-quality experimental PPI data and evolutionary information of a protein pair under prediction. Our experimental results show that DPPI outperforms the state-of-the-art methods on several benchmarks in terms of area under precision-recall curve (auPR), and computationally is more efficient. We also show that DPPI is able to predict homodimeric interactions where other methods fail to work accurately, and the effectiveness of DPPI in specific applications such as predicting cytokine-receptor binding affinities. Predicting protein-protein interactions through sequence-based deep learning): https://github.com/hashemifar/DPPI/. Supplementary data are available at Bioinformatics online.

Abstract PR12: A combined protein-protein interaction and genetic interaction map defines new and critical Kras effectors in non-small cell lung cancer

Abstract Despite intensive study, no drugs in clinical use specifically target KRAS-mutant tumors. Uncharacterized feedback mechanisms and parallel pathways have stymied the treatment of KRAS-mutant tumors with Raf and PI3K inhibitors, and the KRas protein itself does not easily accommodate binding of small-molecule inhibitors. These challenges demand more systematic and quantitative characterization of the physical and genetic relationships between Ras regulators and effectors. To that end, we used tandem affinity purification of Kras, Hras and Nras, their activated alleles and key proteins with known regulatory (GEFs, GAPs) or effector (Raf, RalGDS, RIN1/2) roles in both 293 cells and A549 NSCLC cells to generate a high-confidence protein-protein interaction (PPI) network. This map of 220 proteins and 1,400 physical interactions was used to design an sgRNA library with 10 guides/gene. This library was screened in Cas9-expressing A549 cells and grown for 14 days before analysis for dropout or enhanced representation of sgRNAs. Approximately 120 genes showed positive or negative growth effects. PPIs and genetic interactions (GIs) were cross-referenced with public PPI data and TCGA patient data to assemble a combined physical PPI and genetic map informed by cancer mutations. This map suggests many hypotheses for PPIs critical for growth control. This set was used to construct a sgRNA library covering 120 genes of probable relevance to the Ras pathway with ~60 “safe harbor” control sgRNAs. This library was screened in a two-cassette sgRNA system testing 14K pairwise genetic effects to identify quantitative changes in growth in A549 and H23 NSCLC lines. This screen showed >100 genetic interactions, which in conjunction with PPIs, identify coupling between the Raf/MEK/ERK kinase, Ral and Rap GTPase, RNA processing, and cell adhesion pathways. The screen identified new candidate effector pathways for cell adhesion, RNA processing, Rap GTPase regulation, and protein processing, including the RADIL, RGL, and RIN Kras effectors. Validation focused using the synthetic lethal interactions observed in the sgRNA screen to predict drug combinations showing drug synergy in A549 and H23 cells. Using 11-point dose titrations and isobologram analysis of drug combinations, we see strong synergy among PI3 kinase, Raf, and Erk inhibitors in these cells. Using the recently described Kras G12C inhibitor, expressed in H23 cells, we have validated that sgRNA deletion of the the key Kras effector for specific pathways including cell adhesion (RADIL), growth signaling (RAF), and endocytosis/ macropinocytosis (RIN) are affected and that use of the Kras inhibitor ARS-853 shows much reduced effects on specific Kras effector pathways in cells deleted for these effectors. These systematic data underscore the limitations of our current understanding of Kras-driven cancers, revealing new genetic vulnerabilities and target candidates. This abstract is also being presented as Poster A28. Citation Format: Marcus R. Kelly, Kyuho Han, Nancie Mooney, Edwin Jeng, Kaja Kostyrko, Alejandro Sweet-Cordero, Michael Bassik, Peter K. Jackson. A combined protein-protein interaction and genetic interaction map defines new and critical Kras effectors in non-small cell lung cancer [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr PR12.

A Robust Algorithm for Visualization of Protein Interaction Network

Living systems can be envisioned as a beautiful fabric with weaves of different biological networks such as genetic, protein, transcription factors and metabolic interactions. The weave pattern or the network architecture is a precise predictor of healthy or diseased state of the organism. Network-based approach gives insight into pathogenesis pathway which leads to drug discovery process. It helps researchers and clinicians in grouping together the proteins that interact in functional complexes and pathways thus exploring disease network nodes as potential target for drug discovery. Biological data are often structured in the form of complex interconnected networks such as protein interaction and metabolic networks. Visualization tools help in the visualization of various interactome data. Sheer size, complexity and dynamic nature of networks, and the algorithm responsible for visualization are the main challenges of biological network visualization. Visualization tools are based on certain graph theories and algorithm for visualization. In the present work, directed and undirected graphs are used in the visualization of protein interactions data. To visualize time course behavior of genes, there is need to combine these two graph theories which can interpret biological phenomenon in better way. Further, force-directed algorithm is used for the visualization of protein interaction data. It helps in integrating interactome data and provides better visualization. To overcome this aspect, a holistic approach is required.

Using Two-dimensional Principal Component Analysis and Rotation Forest for Prediction of Protein-Protein Interactions

The interaction among proteins is essential in all life activities, and it is the basis of all the metabolic activities of the cells. By studying the protein-protein interactions (PPIs), people can better interpret the function of protein, decoding the phenomenon of life, especially in the design of new drugs with great practical value. Although many high-throughput techniques have been devised for large-scale detection of PPIs, these methods are still expensive and time-consuming. For this reason, there is a much-needed to develop computational methods for predicting PPIs at the entire proteome scale. In this article, we propose a new approach to predict PPIs using Rotation Forest (RF) classifier combine with matrix-based protein sequence. We apply the Position-Specific Scoring Matrix (PSSM), which contains biological evolution information, to represent protein sequences and extract the features through the two-dimensional Principal Component Analysis (2DPCA) algorithm. The descriptors are then sending to the rotation forest classifier for classification. We obtained 97.43% prediction accuracy with 94.92% sensitivity at the precision of 99.93% when the proposed method was applied to the PPIs data of yeast. To evaluate the performance of the proposed method, we compared it with other methods in the same dataset, and validate it on an independent datasets. The results obtained show that the proposed method is an appropriate and promising method for predicting PPIs.

Prediction of Protein Sub-Mitochondria Locations Using Protein Interaction Networks.

BackgroundPrediction of the protein localization is among the most important issues in the bioinformatics that is used for the prediction of the proteins in the cells and organelles such as mitochondria. In this study, several machine learning algorithms are applied for the prediction of the intracellular protein locations. These algorithms use the features extracted from protein sequences. In contrast, protein interactions have been less investigated.ObjectivesAs protein interactions usually occur in the same or adjacent places, using this feature to find the location would be efficient and impressive. This study did not aim at increasing the total accuracy of the conducted research. The study has focused on the features of the proteins’ interaction and their employment which lead to a higher accuracy.Materials and MethodsIn this study, we have examined the protein interaction network as one of the features for prediction of the protein localization and its effects on the prediction results. In this regards, we have gathered some of the most common features including Amino Acid Composition, Dipeptide Compositions, Pseudo Amino Acid Compositions (PseAAC), Position Specific Scoring Matrix (PSSM), Functional Domain, Gene Ontology information, and the Pair-wise sequence alignment. The results of the classification are compared to the ones using protein interactions. For achieving this goal different machine learning algorithms were tested.ResultsThe best-obtained results of using single feature set obtained using SVM classifier for PseAAC feature. The accuracy of combining all features with PPI data, using the Decision Tree and Random Forest classifiers, was 82.49% and 83.35%, respectively. In another experiment, using just protein interaction data with the different cutting points resulted in obtaining an accuracy of 93.035% for the protein location prediction.ConclusionIn total, it was shown that protein(s) interaction has a significant impact on the prediction of the mitochondrial proteins’ location. This feature can separately distinguish the locations well. Using this feature the accuracy of the results is raised up to 5%.

Prediction of Protein Sub-Mitochondria Locations Using Protein Interaction Networks

Background: Prediction of the protein localization is among the most important issues in the bioinformatics that is used for the prediction of the proteins in the cells and organelles such as mitochondria. In this study, several machine learning algorithms are applied for the prediction of the intracellular protein locations. These algorithms use the features extracted from protein sequences. In contrast, protein interactions have been less investigated.Objectives: As protein interactions usually occur in the same or adjacent places, using this feature to find the location would be efficient and impressive. This study did not aim at increasing the total accuracy of the conducted research. The study has focused on the features of the proteins’ interaction and their employment which lead to a higher accuracy.Materials and Methods: In this study, we have examined the protein interaction network as one of the features for prediction of the protein localization and its effects on the prediction results. In this regards, we have gathered some of the most common features including Amino Acid Composition, Dipeptide Compositions, Pseudo Amino Acid Compositions (PseAAC), Position Specific Scoring Matrix (PSSM), Functional Domain, Gene Ontology information, and the Pair-wise sequence alignment. The results of the classification are compared to the ones using protein interactions. For achieving this goal different machine learning algorithms were tested.Results: The best-obtained results of using single feature set obtained using SVM classifier for PseAAC feature. The accuracy of combining all features with PPI data, using the Decision Tree and Random Forest classifiers, was 82.49% and 83.35%, respectively. In another experiment, using just protein interaction data with the different cutting points resulted in obtaining an accuracy of 93.035% for the protein location prediction.Conclusion: In total, it was shown that protein(s) interaction has a significant impact on the prediction of the mitochondrial proteins’ location. This feature can separately distinguish the locations well. Using this feature the accuracy of the results is raised up to 5%.

A novel temporal protein complexes identification framework based on density–distance and heuristic algorithm

The construction of dynamic protein–protein interaction networks is affected by cell tissue and its biological function, and the identification of protein complexes is important for understanding biological functions. This paper presents a new method named density–distance and heuristic for identifying temporal protein complexes. First, the gene expression data of time course are integrated into the static protein interaction data and a set of time-ordered networks are obtained. Then, the network is integrated with the gene information to calculate the distance between proteins in the protein–protein interaction network. Based on this distance, we have formed a number of clusters and selected the furthest cluster from the other cluster centers as the initial cluster to ensure that nodes with clusters are closest to each other. Finally, a heuristic algorithm is introduced, and the initial cluster is updated in two ways. The experimental results show that the proposed method has better performance compared with the commonly used algorithms; meanwhile, this method has a strong biological significance.

Identification of key pathways and genes in lung carcinogenesis.

The present study aimed to identify key pathways and genes in the pathogenesis of lung cancer. The GSE10072 dataset was downloaded from the Gene Expression Omnibus database. Protein-protein interaction data were collected from Human Protein Reference Database, and 201 pathways were downloaded from the Kyoto Encyclopedia of Genes and Genomes database. Signaling network impact analysis was performed to identify enriched pathways, followed by the construction of a pathway-pathway crosstalk network. Benzopyrene was used to treat normal human lung cells at concentrations of 0.01, 0.1, 1 and 10 µM, and cell viability was measured. Furthermore, growth arrest and DNA damage inducible β (GADD45B), p53, cyclin B, Akt and nuclear factor (NF)-κB protein levels were also measured via western blotting. Impact analysis identified 11 enriched lung cancer-associated KEGG pathways, including ‘complement and coagulation cascades’, ‘ECM-receptor interaction’, ‘P53 signaling pathway’, ‘cell adhesion molecules’ and ‘focal adhesion’. In addition, cell cycle, ‘drug metabolism-cytochrome P450’, ‘metabolic pathways’, ‘pathways in cancer’, ‘focal adhesion’ and ‘antigen processing and presentation’ were central in the pathway-pathway cross-talk network. Furthermore, the upregulated gene GADD45B was associated with three of the pathways, including an activated pathway (‘MAPK signaling pathway’) and two repressed pathways (‘cell cycle’ and ‘P53 pathway’). Western blotting demonstrated that the expression of NF-κB, Akt and GADD45B increased over time in lung cells treated with benzopyrene, whereas the expression levels of cyclin B and P53 decreased. In conclusion, GADD45B may contribute to lung carcinogenesis via affecting the MAPK, P53 signaling and cell cycle pathways.

A Network Hierarchy-Based method for functional module detection in protein–protein interaction networks

In the post-genomic era, one of the important tasks is to identify protein complexes and functional modules from high-throughput protein–protein interaction data, so that we can systematically analyze and understand the molecular functions and biological processes of cells. Although a lot of functional module detection studies have been proposed, how to design correctly and efficiently functional modules detection algorithms is still a challenging and important scientific problem in computational biology. In this paper, we present a novel Network Hierarchy-Based method to detect functional modules in PPI networks (named NHB-FMD). NHB-FMD first constructs the hierarchy tree corresponding to the PPI network and then encodes the tree such that genetic algorithm is employed to obtain the hierarchy tree with Maximum Likelihood. After that functional module partitioning is performed based on it and the best partitioning is selected as the result. Experimental results in the real PPI networks have shown that the proposed algorithm not only significantly outperforms the state-of-the-art methods but also can detect protein modules more effectively and accurately.

Abstract 2269: Transcription factor relationships associated with androgen-deprivation therapy response and metastatic progression in prostate cancer

Abstract Patients with recurrent, aggressive prostate cancer typically undergo androgen-deprivation therapy (ADT), but the benefits are often short-lived, and responses are variable. Failure to respond to ADT invariably leads to metastatic disease, and ultimately death. To investigate differential responses to ADT, we performed whole-transcriptome analysis of 20 patient-matched pre-ADT biopsies and post-ADT prostatectomy specimens, and observed that all patients lost transcriptional signatures indicative of the androgen receptor (AR)-dependent subtype, after treatment. We also identified two subgroups of patients with either a strong or weak transcriptional response to ADT. The strong responders maintained the more aggressive subtype signal, while the weak responders lost expression of these genes and more resembled an AR-suppressed, basal subtype. We generated a strong responder transcriptional network using the PANDA program and integrated expression data from our cohort, protein-protein interaction, and DNA binding motif data. We also leveraged the expression data from a large public dataset of over 800 metastatic and primary samples to construct a metastatic lesion transcriptional network. We identified 20 common transcription factor coordinated groups (TFCGs) associated with both the strong responders and metastatic lesions, including GLI3/GLI2, SOX4/FOXA2/GATA4, ERF/ETV5/ETV3/ELF4, and a TFCG containing JUN, JUNB, JUND, FOS, FOSB, and FOSL1. Many TFCGs in the metastatic network were subsets of larger groups in the strong responders network, implicating these transcription factor associations as potentially critical for both the differential ADT response and metastatic disease progression. Citation Format: Nitya V. Sharma, Kathryn L. Pellegrini, Veronique Ouellet, Felipe O. Giuste, Selvi Ramalingam, Kenneth Watanabe, Eloise Adam-Granger, Lucresse Fossouo, Sungyong You, Michael R. Freeman, Paula Vertino, Karen Conneely, Adeboye O. Osunkoya, Dominique Trudel, Anne-Marie Mes-Masson, John A. Petros, Fred Saad, Carlos S. Moreno. Transcription factor relationships associated with androgen-deprivation therapy response and metastatic progression in prostate cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2269.

Abstract 1313: A combined protein-protein interaction and genetic interaction map defines new and critical Kras effectors in NSCLC

Abstract Despite intensive study, no drugs in clinical use specifically target KRAS-mutant tumors. Uncharacterized feedback mechanisms and parallel pathways have stymied the treatment of KRAS-mutant tumors with Raf and PI3K inhibitors, and the KRas protein itself does not easily accommodate binding of small-molecule inhibitors. These challenges demand more systematic and quantitative characterization of the physical and genetic relationships between Ras regulators and effectors. Thus we used tandem affinity purification of Kras, Hras and Nras, their activated alleles and key proteins with known regulatory (GEFs, GAPs) or effector (Raf, RalGDS) roles in both 293 cells and A549 NSCLC cells to generate a high-confidence protein-protein interaction (PPI) network. This map of 220 proteins and 1,400 physical interactions was used to design an sgRNA library with 10 guides/gene. This library was screened in Cas9-expressing A549 cells and grown for 14 days before analysis for dropout or enhanced representation of sgRNAs. Approximately 120 genes showed positive or negative growth effects. PPIs and genetic interactions (GIs) were cross-referenced with public PPI data and TCGA patient data to assemble a combined physical PPI and genetic map informed by cancer mutations. This map suggests many hypotheses for PPIs critical for growth control. This set was used to construct a sgRNA library covering 120 genes of probable relevance to the Ras pathway with ~60 “safe harbor” control sgRNAs. This library was screened in a two-cassette sgRNA system testing 14K pairwise genetic effects to identify quantitative changes in growth in A549 and H23 NSCLC lines. This screen showed >100 genetic interactions, which in conjunction with PPIs identify coupling between the Raf/MEK/ERK kinase, Ral and Rap GTPase, RNA processing and cell adhesion pathways. The screen identified new candidate effector pathways for cell adhesion, RNA processing, Rap GTPase regulation and protein processing, including the RADIL, RGL and RIN Kras effectors. Validation focused on using the synthetic lethal interactions observed in the sgRNA screen to predict drug combinations showing drug synergy in A549 and H23 cells. Using 11-point dose titrations and isobologram analysis of drug combinations, we see strong synergy among PI3 kinase, Raf and Erk inhibitors in these cells. Using the recently described Kras G12C inhibitor, expressed in H23 cells, we have validated that sgRNA deletion of the the key Kras effector for specific pathway,s including cell adhesion (RADIL), growth signaling (RAF) and endocytosis/ macropinocytosis (RIN), is affected and that use of the Kras inhibitor ARS-853 shows much-reduced effects on specific Kras effector pathways in cells deleted for these effectors. These systematic data underscore the limitations of our current understanding of Kras-driven cancers, revealing new genetic vulnerabilities and target candidates. Citation Format: Peter K. Jackson. A combined protein-protein interaction and genetic interaction map defines new and critical Kras effectors in NSCLC [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 1313.