Francesca Petralia

Proteogenomics connects somatic mutations to signalling in breast cancer

Summary Somatic mutations have been extensively characterized in breast cancer, but the effects of these genetic alterations on the proteomic landscape remain poorly understood. We describe quantitative mass spectrometry-based proteomic and phosphoproteomic analyses of 105 genomically annotated breast cancers of which 77 provided high-quality data. Integrated analyses allowed insights into the somatic cancer genome including the consequences of chromosomal loss, such as the 5q deletion characteristic of basal-like breast cancer. The 5q trans effects were interrogated against the Library of Integrated Network-based Cellular Signatures, thereby connecting CETN3 and SKP1 loss to elevated expression of EGFR, and SKP1 loss also to increased SRC. Global proteomic data confirmed a stromal-enriched group in addition to basal and luminal clusters and pathway analysis of the phosphoproteome identified a G Protein-coupled receptor cluster that was not readily identified at the mRNA level. Besides ERBB2, other amplicon-associated, highly phosphorylated kinases were identified, including CDK12, PAK1, PTK2, RIPK2 and TLK2. We demonstrate that proteogenomic analysis of breast cancer elucidates functional consequences of somatic mutations, narrows candidate nominations for driver genes within large deletions and amplified regions, and identifies therapeutic targets.

Integrative random forest for gene regulatory network inference.

Motivation: Gene regulatory network (GRN) inference based on genomic data is one of the most actively pursued computational biological problems. Because different types of biological data usually provide complementary information regarding the underlying GRN, a model that integrates big data of diverse types is expected to increase both the power and accuracy of GRN inference. Towards this goal, we propose a novel algorithm named iRafNet: integrative random forest for gene regulatory network inference. Results: iRafNet is a flexible, unified integrative framework that allows information from heterogeneous data, such as protein–protein interactions, transcription factor (TF)-DNA-binding, gene knock-down, to be jointly considered for GRN inference. Using test data from the DREAM4 and DREAM5 challenges, we demonstrate that iRafNet outperforms the original random forest based network inference algorithm (GENIE3), and is highly comparable to the community learning approach. We apply iRafNet to construct GRN in Saccharomyces cerevisiae and demonstrate that it improves the performance in predicting TF-target gene regulations and provides additional functional insights to the predicted gene regulations. Availability and implementation: The R code of iRafNet implementation and a tutorial are available at: http://research.mssm.edu/tulab/software/irafnet.html Contact: zhidong.tu@mssm.edu Supplementary information: Supplementary data are available at Bioinformatics online.

New Method for Joint Network Analysis Reveals Common and Different Coexpression Patterns among Genes and Proteins in Breast Cancer

We focus on characterizing common and different coexpression patterns among RNAs and proteins in breast cancer tumors. To address this problem, we introduce Joint Random Forest (JRF), a novel nonparametric algorithm to simultaneously estimate multiple coexpression networks by effectively borrowing information across protein and gene expression data. The performance of JRF was evaluated through extensive simulation studies using different network topologies and data distribution functions. Advantages of JRF over other algorithms that estimate class-specific networks separately were observed across all simulation settings. JRF also outperformed a competing method based on Gaussian graphic models. We then applied JRF to simultaneously construct gene and protein coexpression networks based on protein and RNAseq data from CPTAC-TCGA breast cancer study. We identified interesting common and differential coexpression patterns among genes and proteins. This information can help to cast light on the potential disease mechanisms of breast cancer.

Discover the network mechanisms underlying the connections between aging and age-related diseases.

Although our knowledge of aging has greatly expanded in the past decades, it remains elusive why and how aging contributes to the development of age-related diseases (ARDs). In particular, a global mechanistic understanding of the connections between aging and ARDs is yet to be established. We rely on a network modelling named “GeroNet” to study the connections between aging and more than a hundred diseases. By evaluating topological connections between aging genes and disease genes in over three thousand subnetworks corresponding to various biological processes, we show that aging has stronger connections with ARD genes compared to non-ARD genes in subnetworks corresponding to “response to decreased oxygen levels”, “insulin signalling pathway”, “cell cycle”, etc. Based on subnetwork connectivity, we can correctly “predict” if a disease is age-related and prioritize the biological processes that are involved in connecting to multiple ARDs. Using Alzheimer’s disease (AD) as an example, GeroNet identifies meaningful genes that may play key roles in connecting aging and ARDs. The top modules identified by GeroNet in AD significantly overlap with modules identified from a large scale AD brain gene expression experiment, supporting that GeroNet indeed reveals the underlying biological processes involved in the disease.

A new method for constructing tumor specific gene co-expression networks based on samples with tumor purity heterogeneity

Motivation Tumor tissue samples often contain an unknown fraction of stromal cells. This problem is widely known as tumor purity heterogeneity (TPH) was recently recognized as a severe issue in omics studies. Specifically, if TPH is ignored when inferring co‐expression networks, edges are likely to be estimated among genes with mean shift between non‐tumor‐ and tumor cells rather than among gene pairs interacting with each other in tumor cells. To address this issue, we propose Tumor Specific Net (TSNet), a new method which constructs tumor‐cell specific gene/protein co‐expression networks based on gene/protein expression profiles of tumor tissues. TSNet treats the observed expression profile as a mixture of expressions from different cell types and explicitly models tumor purity percentage in each tumor sample. Results Using extensive synthetic data experiments, we demonstrate that TSNet outperforms a standard graphical model which does not account for TPH. We then apply TSNet to estimate tumor specific gene co‐expression networks based on TCGA ovarian cancer RNAseq data. We identify novel co‐expression modules and hub structure specific to tumor cells. Availability and implementation R codes can be found at https://github.com/petraf01/TSNet.

A new method to study the change of miRNA–mRNA interactions due to environmental exposures

Motivation: Integrative approaches characterizing the interactions among different types of biological molecules have been demonstrated to be useful for revealing informative biological mechanisms. One such example is the interaction between microRNA (miRNA) and messenger RNA (mRNA), whose deregulation may be sensitive to environmental insult leading to altered phenotypes. The goal of this work is to develop an effective data integration method to characterize deregulation between miRNA and mRNA due to environmental toxicant exposures. We will use data from an animal experiment designed to investigate the effect of low‐dose environmental chemical exposure on normal mammary gland development in rats to motivate and evaluate the proposed method. Results: We propose a new network approach—integrative Joint Random Forest (iJRF), which characterizes the regulatory system between miRNAs and mRNAs using a network model. iJRF is designed to work under the high‐dimension low‐sample‐size regime, and can borrow information across different treatment conditions to achieve more accurate network inference. It also effectively takes into account prior information of miRNA‐mRNA regulatory relationships from existing databases. When iJRF is applied to the data from the environmental chemical exposure study, we detected a few important miRNAs that regulated a large number of mRNAs in the control group but not in the exposed groups, suggesting the disruption of miRNA activity due to chemical exposure. Effects of chemical exposure on two affected miRNAs were further validated using breast cancer human cell lines. Availability and implementation: R package iJRF is available at CRAN. Contacts: pei.wang@mssm.edu or susan.teitelbaum@mssm.edu Supplementary information: Supplementary data are available at Bioinformatics online.

A next generation sequencing based approach to identify extracellular vesicle mediated mRNA transfers between cells

BackgroundExosomes and other extracellular vesicles (EVs) have emerged as an important mechanism of cell-to-cell communication. However, previous studies either did not fully resolve what genetic materials were shuttled by exosomes or only focused on a specific set of miRNAs and mRNAs. A more systematic method is required to identify the genetic materials that are potentially transferred during cell-to-cell communication through EVs in an unbiased manner.ResultsIn this work, we present a novel next generation of sequencing (NGS) based approach to identify EV mediated mRNA exchanges between co-cultured adipocyte and macrophage cells. We performed molecular and genomic profiling and jointly considered data from RNA sequencing (RNA-seq) and genotyping to track the “sequence varying mRNAs” transferred between cells. We identified 8 mRNAs being transferred from macrophages to adipocytes and 21 mRNAs being transferred in the opposite direction. These mRNAs represented biological functions including extracellular matrix, cell adhesion, glycoprotein, and signal peptides.ConclusionsOur study sheds new light on EV mediated RNA communications between adipocyte and macrophage cells, which may play a significant role in developing insulin resistance in diabetic patients. This work establishes a new method that is applicable to examining genetic material exchanges in many cellular systems and has the potential to be extended to in vivo studies as well.

Abstract IA29: Proteogenomic and phosphoproteomic analysis of breast cancer

The genetic landscape of human breast cancer has been well defined in The Cancer Genome Atlas (TCGA) project. Mass spectrometry (MS)-based global proteome and phosphoproteome analyses provide a complementary, orthogonal approach to genomic studies to further improve the molecular taxonomy and biological understanding of breast cancer. We analyzed human breast cancer samples that had previously undergone comprehensive genomic and reversed phase protein array (RPPA) characterization by TCGA. Tumor samples were analyzed by global shotgun proteomics and phosphoproteomics at an unprecedented coverage of >11,000 quantified proteins and >27,000 phosphorylation sites for each tumor. We verified the translation of hundreds of genomically characterized single nucleotide and splice junction variants at the protein level. The correlation of mRNA to protein abundance was significant for 6,135 out of 9,302 protein/mRNA pairs, but differed amongst protein classes. Genes that did not correlate on the protein/mRNA level included components of basic cellular machineries such as the ribosome, RNA polymerase and spliceosome, as well as those involved in processes regulated by proteolysis. Hierarchical clustering yielded three major clusters in both the proteome and the phosphoproteome data: basal-enriched, luminal-enriched and stroma-enriched groups, the last also enriched for what have been previously designated “reactive-type” tumors by RPPA. Our deep proteome analysis promoted new insights including the consequences of chromosomal loss, such as the 5q deletion characteristic of basal-like breast cancer. The 5q trans effects were interrogated using the Library of Integrated Network-based Cellular Signatures. Theses analyses connected the 5q genes CETN3 and SKP1 to elevated expression of EGFR, and SKP1 also to SRC. Differential phosphopeptide analyses, integrated with activity maps derived from knock-in mutated cell lines, identified multiple novel downstream effects of PIK3CA and TP53 mutation. Besides ERBB2, other amplicon-associated, highly phosphorylated kinases were identified, including CDK12, PAK1, PTK2, RIPK2 and TLK2. These and other examples demonstrate that proteogenomic analysis of breast cancer elucidates functional consequences of somatic mutations, narrows candidate nominations for driver genes within large deletions and amplified regions, and identifies potential therapeutic targets. Citation Format: Philipp Mertins, DR Mani, Kelly Ruggles, Michael Gillette, Karl Clauser, Pei Wang, Xianlong Wang, Jana Qiao, Song Cao, Francesca Petralia, Filip Mundt, Zhidong Tu, Jonathan Lei, Michael Gatza, Matthew Wilkerson, Charles Perou, Venkata Yellapantula, Kuan-lin Huang, Chenwei Lin, Michael McLellan, Ping Yan, Sherri Davies, Reid Townsend, Steven Skates, Jing Wang, Bing Zhang, Christopher Kinsinger, Mehdi Mesri, Henry Rodriguez, Li Ding, Amanda Paulovich, David Fenyo, Matthew Ellis, Steven Carr, NCI CPTAC. Proteogenomic and phosphoproteomic analysis of breast cancer. [abstract]. In: Proceedings of the AACR Special Conference on Advances in Breast Cancer Research; Oct 17-20, 2015; Bellevue, WA. Philadelphia (PA): AACR; Mol Cancer Res 2016;14(2_Suppl):Abstract nr IA29.

Abstract 798: Effects of low-dose environmental chemicals on the mammary transcriptome at critical windows of development in a rodent model

Exposure to environmental chemicals, including those commonly found in personal care products has been linked to mammary cancer at high doses using animal models. Their effects at low doses comparable to human exposure, especially during critical windows of development remain poorly understood. We investigated the effects of three prevalent environmental chemicals - diethyl phthalate (DEP), methyl paraben (MPB), triclosan (TCS) - and their mixture (MIX) on the transcriptome of normally developing mammary at low doses mimicking human exposure. Using a female Sprague-Dawley rat model, we targeted four early developmental exposure windows - prenatal, neonatal, prepubertal and pubertal, as well as continuous exposure from birth to adulthood (both parous and nulliparous). Control rats were exposed to vehicle only. All exposures were by oral gavage. Whole-transcriptomes of mammary glands were profiled by Affymetrix rat gene arrays. Differentially expressed genes were identified by linear models. Despite dynamic transcriptome changes in the normal developing mammary, exposure to environmental chemicals induced detectable gene expression changes in a window-specific fashion. We discovered that puberty represented a window of heightened sensitivity to MPB and DEP exposure with 341 and 175 altered genes relative to controls, respectively (false discovery rate (FDR) Citation Format: Kalpana Gopalakrishnan, Francesca Petralia, Pei Wang, Fabiana Manservisi, Laura Falcioni, Luciano Bua, Fiorella Belpoggi, Luca Lambertini, James Wetmur, Susan Teitelbaum, Jia Chen, Vasily Aushev. Effects of low-dose environmental chemicals on the mammary transcriptome at critical windows of development in a rodent model. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 798.

Abstract 782: A network approach to study the effect of chemical exposures on gene regulatory system in rats

Exposure to environmental chemicals during early development may increase the risk of developing breast cancer later in life. In this context, we are interested in characterizing which microRNA (miRNA) and mRNA expressions change in a coherent manner across the lifespan, and whether the co-expression pattern is affected by environmental exposures. miRNAs contribute to tumor progression via the regulation of post-transcriptional gene expressions. Thus, information on different interaction patterns among miRNAs and mRNAs measured in mammary tissues from chemical exposed vs. non-exposed rats can cast light on how chemical exposures may alter mammary gland development. Specifically, we consider three common environmental chemicals: diethyl phthalate (DEP), methyl paraben (MPB) and triclosan (TCS). Female Sprague-Dawley rats were treated with these chemicals at four windows of susceptibility (prenatal, neonatal, prepubertal and pubertal) respectively with oral doses shown to produce urinary metabolite levels similar to those measured in US population. We implemented a new algorithm, Joint Random Forest (JRF), for simultaneous estimation of multiple related networks to characterize co-expression patterns among mRNAs and miRNAs. JRF is designed to borrow information across different treatment conditions, so that regulatory relationships shared across conditions can be detected with increased power, while those specific to each condition can be detected with fewer false positives. We focused on 1403 mRNAs and 283 miRNAs with larger variability across rats, and derived four co-expression networks of these mRNAs/miRNAs for each environmental chemical treatment plus a control group. Overall we observed a substantial loss of connectivity in networks of chemical exposed groups (DEP: 1813 edges, MPB: 1539 edges and TCS: 1013 edges) compared to that of the control group (2641 edges). Interestingly, despite the overall loss of connectivity in networks of chemical exposed groups, some microRNAs such as rno-miR-126b (MIMAT0017843) and rno-miR-3596b (MIMAT0017871) showed many connecting edges in networks of chemical exposed groups but zero in that of the control group. In particular, rno-miR-3596b is a member of the Let-7 family which is well known to regulate self-renewal and tumorigenicity in breast cancer cells. Findings like this can lead to better understanding of how chemical exposures could alter gene regulatory activities. Our study also demonstrates the great potential of using JRF to investigate changes in gene regulatory system across different conditions. Citation Format: Francesca Petralia, Vasily Aushev, Kalpana Gopalakrishnan, Susan Teitelbaum, Jia Chen, Pei Wang. A network approach to study the effect of chemical exposures on gene regulatory system in rats. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 782.