SEARCH · Search bioRxiv

Results for “Cancer Biology”

Search indexed bioRxiv preprints in genomics, neuroscience, cell biology and bioinformatics. Read source abstracts and check manuscript versions; preprints are not peer reviewed.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,495 records · Page 83Linked to original sources

bioRxiv2017

Increased glutaminolytic flux and activation of mitochondrial metabolism by BCL2 hyperactivity in lymphoma

B-cell lymphoma 2 (BCL2) is an important apoptosis regulator during developmental and pathological states, and its overexpression is a key feature of several malignancies. Genomic data from The Cancer Genome Atlas (TCGA) reveals significant somatic copy number amplification, overexpression, and/or elevated protein activity of BCL2 in 50 % of diffuse large B-cell lymphoma (DLBC) patients. While its canonical role in mitochondria-directed apoptosis is well established, the effect of BCL2 on transcriptional and metabolic networks remains elusive. Using an established lymphocytic pro-B-cell line overexpressing BCL2, we identified dysregulated transcriptional and metabolic networks by transcriptomic profiling arrays. Elevated BCL2 levels affect transcription factor complexes and mitogenic programs of NF-{kappa}B/REL, HIF1A/ARNT, AP1, E2F, and STAT factors. Using stable isotope-assisted metabolic flux measurements we quantify that elevated BCL2 expression increases carbon utilization boosting cellular proliferation. Tumorigenic overexpression of BCL2 significantly increases glycolytic flux, glutaminolysis, and anaplerotic flux into the TCA cycle. At the same time, the mitochondrial acetyl-CoA pool is separated from the glycolytic one by inactivating the pyruvate dehydrogenase complex via transcriptional regulation of pyruvate dehydrogenase kinase (PDK3). As compensatory fuel, mitochondrial TCA cycle metabolism is supported by asparagine synthase (ASNS) and oxidative glutaminolysis creating targets for small molecule inhibition of glutaminase. Lymphoma cells overexpressing BCL2 contained more mitochondrial mass and were more sensitive to L-glutamine deprivation and glutaminase inhibition. Cells overexpressing a mutant BCL2 G145E, which is incapable of binding BH domain members, failed to increase proliferation, glycolysis, or glutaminolysis. Taken together, the oncogene BCL2 has the ability to ramp up a metabolic phenotype supporting proliferation independent of its anti-apoptotic role. The cellular model of BCL2 activation supports NF-KB-positive subtypes of DLBC and identifies metabolic bottlenecks with dependency on anaplerotic flux as an actionable BCL2 effector network in cancer.

systems biology↗

bioRxiv2016

Modeling cumulative biological phenomena with Suppes-Bayes causal networks

Several diseases related to cell proliferation are characterized by the accumulation of somatic DNA changes, with respect to wildtype conditions. Cancer and HIV are two common examples of such diseases, where the mutational load in the cancerous/viral population increases over time. In these cases, selective pres sures are often observed along with competition, co-operation and parasitism among distinct cellular clones. Recently, we presented a mathematical framework to model these phenomena, based on a combination of Bayesian inference and Suppes theory of probabilistic causation, depicted in graphical structures dubbed Suppes-Bayes Causal Networks (SBCNs). SBCNs are generative probabilistic graphical models that recapitulate the potential ordering of accumulation of such DNA changes during the progression of the disease. Such models can be inferred from data by exploiting likelihood-based model-selection strategies with regularization. In this paper we discuss the theoretical foun dations of our approach and we investigate in depth the influence on the model-selection task of: (i) the poset based on Suppes theory and (ii) different regulariza tion strategies. Furthermore, we provide an example of application of our framework to HIV genetic data highlighting the valuable insights provided by the inferred SBCN.

Bioinformatics↗

bioRxiv2018

Network-based genetic profiling, and therapeutic target identification of Thyroid Cancer

Pathogenic mechanisms that underlie malignant follicular thyroid carcinoma (FTC) development are poorly understood. To identify key genes and pathways driving malignant behaviour we employed a system biology-based integrative analyses comparing FTC transcriptomes with a similar but benign lesion, follicular thyroid adenoma (FTA). We identified differentially expressed genes (DEGs) in microarray gene expression datasets (n=52) of FTCs and FTA tissues. Pathway analyses of DEGs using gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) resources revealed significant pathways, and pathway hub genes using protein-protein interactions (PPI). We identified 598 DEGs (relative to FTAs) in FTCs and 12 significant pathways with altered expression in FTC. 10 GO groups were significantly connected with FTC-high expression DEGs and 80 with low-FTC expression. PPI analysis identified 12 potential hub genes based on degree and betweenness centrality. Moreover, 10 transcription factors (TFs) were identified that may underlie DEG expression as well as a number of microRNA (miRNAs). Thus, we identified DEGs, pathways, TFs and miRNAs that reflect molecular mechanisms differing between FTC and benign FTA. These may constitute biomarkers that distinguish these lesions and, given the similarities and common origin of the lesions, they may also be indicators of malignant progression potential.

bioinformatics↗

bioRxiv2018

FRACTIONAL DELETION OF COMPOUND KUSHEN INJECTION, A NATURAL COMPOUND MIXTURE, INDICATES CYTOKINE SIGNALING PATHWAYS ARE CRITICAL FOR ITS PERTURBATION OF THE CELL CYCLE.

We have used computational and experimental biology approaches to identify candidate mechanisms of action of a traditional Chinese medicine. Compound Kushen Injection (CKI), in a breast cancer cell line in which CKI causes apoptosis. Because CKI is a complex mixture of plant secondary metabolites, we used a high-performance liquid chromatography (HPLC) fractionation and reconstitution approach to define chemical fractions required for CKI to induce apoptosis in MDA-MB-231 cells. Our initial fractionation separated major from minor compounds, and showed that the major compounds accounted for little of the activity of CKI. By systematically perturbing the major compounds in CKI we found that removal of no single major compound could alter the effect of CKI on cell viability and apoptosis. However, simultaneous removal of two major compounds identified oxymatrine and oxysophocarpine as critical compounds with respect to CKI activity. We then used RNA sequencing and transcriptome analysis to correlate compound removal with gene expression and phenotype data. We determined that many compounds in CKI are required for its effectiveness in triggering apoptosis but that significant modulation of its activity is conferred by a small number of compounds. In conclusion, CKI may be typical of many plant based extracts that contain many compounds in that no single compound is responsible for all of the bioactivity of the mixture and that many compounds interact in a complex fashion to influence a network containing many targets.

pharmacology and toxicology↗

bioRxiv2016

Comprehensive discovery of subsample gene expression components by information explanation: therapeutic implications in cancer

BackgroundDe novo inference of clinically relevant gene function relationships from tumor RNA-seq remains a challenging task. Current methods typically either partition patient samples into a few subtypes or rely upon analysis of pairwise gene correlations (co-expression) that will miss some groups in noisy data. Leveraging higher dimensional information can be expected to increase the power to discern targetable pathways, but this is commonly thought to be an intractable computational problem.\n\nMethodsIn this work we adapt a recently developed machine learning algorithm, CorEx, that efficiently optimizes over multivariate mutual information for sensitive detection of complex gene relationships. The algorithm can be iteratively applied to generate a hierarchy of latent factors. Patients are stratified relative to each factor and combinatoric survival analyses are performed and interpreted in the context of biological function annotations and protein network interactions that might be utilized to match patients to multiple therapies.\n\nResultsAnalysis of ovarian tumor RNA-seq samples demonstrates the algorithms power to infer well over one hundred biologically interpretable gene cohorts, several times more than standard methods such as hierarchical clustering and k-means. The CorEx factor hierarchy is also informative, with related but distinct gene clusters grouped by upper nodes. Some latent factors correlate with patient survival, including one for a pathway connected with the epithelial-mesenchymal transition in breast cancer that is regulated by a potentially druggable microRNA. Further, combinations of factors lead to a synergistic survival advantage in some cases.\n\nConclusionsIn contrast to studies that attempt to partition patients into a small number of subtypes (typically 4 or fewer) for treatment purposes, our approach utilizes subgroup information for combinatoric transcriptional phenotyping. Considering only the 66 gene expression groups that are both found to have significant Gene Ontology enrichment and are small enough to indicate specific drug targets implies a computational phenotype for ovarian cancer that allows for 366 possible patient profiles, enabling truly personalized treatment. The findings here demonstrate a new technique that sheds light on the complexity of gene expression dependencies in tumors and could eventually enable the use of patient RNA-seq profiles for selection of personalized and effective cancer treatments.

Systems Biology↗

bioRxiv2016

Knowledge Formalization and High-Throughput Data Visualization Using Signaling Network Maps

Generation and usage of high-quality molecular signalling network maps can be augmented by standardising notations, establishing curation workflows and application of computational biology methods to exploit the knowledge contained in the maps. In this manuscript, we summarize the major aims and challenges of assembling information in the form of comprehensive maps of molecular interactions. Mainly, we share our experience gained while creating the Atlas of Cancer Signalling Network. In the step-by-step procedure, we describe the map construction process and suggest solutions for map complexity management by introducing a hierarchical modular map structure. In addition, we describe the NaviCell platform, a computational technology using Google Maps API to explore comprehensive molecular maps similar to geographical maps, and explain the advantages of semantic zooming principles for map navigation. We also provide the outline to prepare signalling network maps for navigation using the NaviCell platform. Finally, several examples of cancer high-throughput data analysis and visualization in the context of comprehensive signalling maps are presented.

systems biology↗

bioRxiv2014

Musashi proteins are post-transcriptional regulators of the epithelial-luminal cell state

SummaryThe conserved Musashi (Msi) family of RNA binding proteins are expressed in stem/progenitor and cancer cells, but mostly absent from differentiated cells, consistent with a role in cell state regulation. We found that Msi genes are rarely mutated but frequently overexpressed in human cancers, and associated with an epithelial-luminal cell state. Using ribosome footprint profiling and RNA-seq analysis of genetic mouse models in neuronal and mammary cell types, we found that Msis regulate translation of genes implicated in epithelial cell biology and epithelial-to-mesenchymal transition (EMT) and promote an epithelial splicing pattern. Overexpression of Msi proteins inhibited translation of genes required for EMT, including Jagged1, and repressed EMT in cell culture and in mammary gland in vivo, while knockdown in epithelial cancer cells led to loss of epithelial identity. Our results show that mammalian Msi proteins contribute to an epithelial gene expression program and promote an epithelial-luminal state in both neural and breast cell types.\n\nHighlightsO_LIMsi proteins bind UAG motifs in vitro and in 3 UTRs of mRNAs\nC_LIO_LIMsi proteins are markers of epithelial state in brain and breast tumors, and cell lines\nC_LIO_LIThe Notch regulator Jag1 mRNA is bound and translationally repressed by Msi\nC_LIO_LIMsi overexpression represses EMT in the mammary gland in vivo\nC_LI

Genomics↗

bioRxiv2015

Agent-based modeling of cancer stem cell driven solid tumor growth

Computational modeling of tumor growth has become an invaluable tool to simulate complex cell-cell interactions and emerging population-level dynamics. Agent-based models are commonly used to describe the behavior and interaction of individual cells in different environments. Behavioral rules can be informed and calibrated by in vitro assays, and emerging population-level dynamics may be validated with both in vitro and in vivo experiments. Here, we describe the design and implementation of a lattice-based agent-based model of cancer stem cell driven tumor growth.

Systems Biology↗

bioRxiv2016

Modeling and Analysis of Hormone and Mitogenic Signal Integration in Prostate Cancer

Prostate cancer is the most common cancer in men and the second leading cause of cancer related death in the United States. Androgens, such as testosterone, are required for androgen dependent prostate cancer (ADPC) growth. Androgen ablation in combination with radiation or chemotherapy remains the primary non-surgical treatment for ADPC. However, androgen ablation typically fails to permanently arrest cancer progression, often resulting in castration resistant prostate cancer (CRPC). In this study, we analyzed a population of mathematical models that described the integration of androgen and mitogenic signaling in androgen dependent and independent prostate cancer. An ensemble of model parameters was estimated from 43 studies of signaling in androgen dependent and resistant LNCaP cell lines. The model population was then validated by comparing simulations with an additional 33 data sets from LNCaP cell lines and clinical trials. Analysis of the model population suggested that simultaneously targeting the PI3K and MAPK pathways in addition to anti-androgen therapies could be an effective treatment for CRPC. We tested this hypothesis in both ADPC LNCaP cell lines and LNCaP derived CRPC C4-2 cells using three inhibitors: the androgen receptor inhibitor MDV3100 (enzalutamide), the Raf kinase inhibitor sorafenib, and the PI3K inhibitor LY294002. Consistent with model predictions, cell viability decreased at 72 hrs in the dual and triple inhibition cases in both the LNCaP and C4-2 cell lines, compared to treatment with any single inhibitor. Taken together, this study suggested that crosstalk between the androgen and mitogenic signaling axes led to robustness of CRPC to any single inhibitor. Model analysis predicted potentially efficacious target combinations which were confirmed by experimental studies in multiple cell lines, thereby illustrating the potentially important role that mathematical modeling can play in cancer.

Systems Biology↗

bioRxiv2016

Temporal regulation of epithelium formation

Abstract/SummaryTo establish the animal body plan, embryos link the external epidermis to the internal digestive tract. In Caenorhabditis elegans, this linkage is achieved by the Arcade Cells, which form an epithelial bridge between the foregut and epidermis, but little is known about how development of these three epithelia is coordinated temporally. The Arcade Cell epithelium is generated after the epidermis and digestive tract epithelia have matured, ensuring that both organs can withstand the mechanical stress of embryo elongation; mis-timing of epithelium formation leads to defects in morphogenesis. Here, we report that temporal regulation of the Arcade Cell epithelium is mediated by the pioneer transcription factor PHA-4/FoxA, the cytoskeletal regulator ZEN-4/MKLP and the polarity protein PAR-6. We find that PHA-4 activates expression of a broad cohort of epithelial genes. However, accumulation of protein is delayed by ZEN-4, acting in concert with its partner CYK-4/MgcRacGAP. Finally, PAR-6 localizes factors within adherens junctions and at the apical surface, leading to Arcade Cell polarity. The results reveal that the timing of a landmark event during embryonic morphogenesis is mediated by the concerted action of four proteins that delay the formation of an epithelial bridge. In addition, we find that FoxA associates with many epithelial genes in mammals, suggesting that regulation of epithelial identity may be a conserved feature of FoxA factors and a contributor to FoxA function in development and cancer.

Cell Biology↗

bioRxiv2017

Dynamics of human telomerase recruitment depend on template-telomere base-pairing

The reverse transcriptase telomerase adds telomeric repeats to chromosome ends to counteract telomere shortening and thereby assures genomic stability in dividing human cells. Key variables in telomere homeostasis are the frequency with which telomerase engages the chromosome end and the number of telomeric repeats it adds during each association event. To study telomere elongation in vivo we have established a live-cell imaging assay to track individual telomerase RNPs in HeLa cells. Using this assay and the drug imetelstat, which is a competitive inhibitor of telomeric DNA binding, we demonstrate that stable association of telomerase with the single-stranded overhang of the chromosome end requires telomerase-DNA base-pairing. Furthermore, we show that telomerase processivity contributes to telomere elongation in vivo. Together, these findings provide new insight into the dynamics of telomerase recruitment and the importance of processivity in maintaining telomere length in human cancer cells.

cell biology↗

bioRxiv2018

Infectious Entry of Merkel Cell Polyomavirus

Merkel Cell Polyomavirus (MCPyV) is a small, non-enveloped tumor virus associated with an aggressive form of skin cancer, the Merkel cell carcinoma (MCC). MCPyV infections are highly prevalent in the human population with MCPyV virions being continuously shed from human skin. However, the precise host cell tropism(s) of MCPyV remains unclear: MCPyV is able to replicate within a subset of dermal fibroblasts, but MCPyV DNA has also been detected in a variety of other tissues. However, MCPyV appears different from other polyomaviruses as it requires sulfated polysaccharides such as heparan sulfates and/or chondroitin sulfates for initial attachment. Like other polyomaviruses, MCPyV engages sialic acid as a (co-receptor). To explore the infectious entry process of MCPyV, we analyzed the cell biological determinants of MCPyV entry into A549 cells, a highly transducible lung carcinoma cell line, in comparison to well-studied simian virus 40 and a number of other viruses. Our results indicate that MCPyV enters cells via caveolar/lipid raft-mediated endocytosis but not macropinocytosis, clathrin-mediated endocytosis or glycosphingolipid-enriched carriers. The viruses internalized in small endocytic pits that led the virus to endosomes and from there to the endoplasmic reticulum (ER). Similar to other polyomaviruses, trafficking required microtubular transport, acidification of endosomes, and a functional redox environment. To our surprise, the virus was found to acquire a membrane envelope within endosomes, a phenomenon not reported for other viruses. Only minor amounts of viruses reached the ER, while the majority was retained in endosomal compartments suggesting that endosome-to-ER trafficking is a bottleneck during infectious entry.\n\nImportanceMCPyV is the first polyomavirus directly implicated in the development of an aggressive human cancer, the Merkel Cell Carcinoma (MCC). Although MCPyV is constantly shed from healthy skin, MCC incidence increases among aging and immunocompromised individuals. To date, the events connecting initial MCPyV infection and subsequent transformation still remain elusive. MCPyV differs from other known polyomaviruses concerning its cell tropism, entry receptor requirements, and infection kinetics. In this study, we examined the cellular requirements for endocytic entry as well as the subcellular localization of incoming virus particles. A thorough understanding of the determinants of the infectious entry pathway and the specific biological niche will benefit prevention of virus-derived cancers such as MCC.

microbiology↗

bioRxiv2018

Haplotype-resolved and integrated genome analysis of ENCODE cell line HepG2

The HepG2 cancer cell line is one of the most widely-used biomedical research and one of the main cell lines of ENCODE. Vast numbers of functional genomics and epigenomics datasets have been produced to characterize its biology. However, the correct interpretation such data requires an understanding of the cell lines genome sequence and genome structure. Using a variety of sequencing and analysis methods, we identified a wide spectrum of HepG2 genome characteristics: copy numbers of chromosomal segments, SNVs and Indels (corrected for aneuploidy), phased haplotypes extending to entire chromosome arms, loss of heterozygosity, retrotransposon insertions, structural variants (SVs) including complex and somatic genomic rearrangements. We also identified allele-specific expression and DNA methylation genome-wide and assembled an allele-specific CRISPR/Cas9 targeting map.\n\nSIGNIFICANCEHaplotype-resolved and comprehensive whole-genome analysis of a widely-used cell line for cancer research and ENCODE, HepG2, serves as an essential resource for unlocking complex cancer gene regulation using a genome-integrated framework and also provides genomic context for the analysis of ~1,000 functional datasets to date on ENCODE for biological discovery. We also demonstrate how deeper insights into genomic regulatory complexity are gained by adopting a genome-integrated framework.

genomics↗

bioRxiv2017

confFuse: high-confidence fusion gene detection across tumor entities

BackgroundFusion genes play an important role in the tumorigenesis of many cancers. Next-generation sequencing (NGS) technologies have been successfully applied in fusion gene detection for the last several years, and a number of NGS-based tools have been developed for identifying fusion genes during this period. Most fusion gene detection tools based on RNA-seq data report a large number of candidates (mostly false positives), making it hard to prioritize candidates for experimental validation and further analysis. Selection of reliable fusion genes for downstream analysis becomes very important in cancer research. We therefore developed confFuse, a scoring algorithm to reliably select high-confidence fusion genes which are likely to be biologically relevant.\n\nResultsConfFuse takes multiple parameters into account in order to assign each fusion candidate a confidence score, of which score [≥]8 indicates high-confidence fusion gene predictions. These parameters were manually curated based on our experience and on certain structural motifs of fusion genes. Compared with alternative tools, based on 96 published RNA-seq samples from different tumor entities, our method can significantly reduce the number of fusion candidates (301 high-confidence from 8,083 total predicted fusion genes) and keep high detection accuracy (recovery rate 85.7%). Validation of 18 novel, high-confidence fusions detected in three breast tumor samples resulted in a 100% validation rate.\n\nConclusionsConfFuse is a novel downstream filtering method that allows selection of highly reliable fusion gene candidates for further downstream analysis and experimental validations. confFuse is available at https://github.com/Zhiqin-HUANG/confFuse.

bioinformatics↗

bioRxiv2018

Metabolic and signalling network map integration: application to cross-talk studies and omics data analysis in cancer

BackgroundThe interplay between metabolic processes and signalling pathways remains poorly understood. Global, detailed and comprehensive reconstructions of human metabolism and signalling pathways exist in the form of molecular maps, but they have never been integrated together. We aim at filling in this gap by creating an integrated resource of both signalling and metabolic pathways allowing a visual exploration of multi-level omics data and study of cross-regulatory circuits between these processes in health and in disease.\n\nResultsWe combined two comprehensive manually curated network maps. Atlas of Cancer Signalling Network (ACSN), containing mechanisms frequently implicated in cancer; and ReconMap 2.0, a comprehensive reconstruction of human metabolic network. We linked ACSN and ReconMap 2.0 maps via common players and represented the two maps as interconnected layers using the NaviCell platform for maps exploration. In addition, proteins catalysing metabolic reactions in ReconMap 2.0 were not previously visually represented on the map canvas. This precluded visualisation of omics data in the context of ReconMap 2.0. We suggested a solution for displaying protein nodes on the ReconMap 2.0 map in the vicinity of the corresponding reaction or process nodes. This permits multi-omics data visualisation in the context of both map layers. Exploration and shuttling between the two map layers is possible using Google Maps-like features of NaviCell. The integrated ACSN-ReconMap 2.0 resource is accessible online and allows data visualisation through various modes such as markers, heat maps, bar-plots, glyphs and map staining. The integrated resource was applied for comparison of immunoreactive and proliferative ovarian cancer subtypes using transcriptomic, copy number and mutation multi-omics data. A certain number of metabolic and signalling processes specifically deregulated in each of the ovarian cancer sub-types were identified.\n\nConclusionsAs knowledge evolves and new omics data becomes more heterogeneous, gathering together existing domains of biology under common platforms is essential. We believe that an integrated ACSN-ReconMap 2.0 resource will help in understanding various disease mechanisms and discovery of new interactions at the intersection of cell signalling and metabolism. In addition, the successful integration of metabolic and signalling networks allows broader systems biology approach application for data interpretation and retrieval of intervention points to tackle simultaneously the key players coordinating signalling and metabolism in human diseases.

systems biology↗

bioRxiv2018

Estimating heterogeneous treatment effects by balancing heterogeneity and fitness

Estimating heterogeneous treatment effects is an important problem in many medical and biological applications since treatments may have different effects on the prognoses of different patients. Recently, several recursive partitioning methods have been proposed to identify the subgroups that with different responds to a treatment, and they rely on a fitness criterion to minimize the error between the estimated treatment effects and the unobservable true effects. In this paper, we propose that a heterogeneity criterion, which maximizes the differences of treatment effects among the subgroups, also needs to be considered. Moreover, we show that better performances can be achieved when the fitness and the heterogeneous criteria are considered simultaneously. Selecting the optimal splitting points then becomes a multi-objective problem; however, a solution that achieves optimal in both aspects are often not available. To solve this problem, we propose a multi-objective splitting procedure to balance both criteria. The proposed procedure is computationally efficient and fits naturally into the existing recursive partitioning framework. Experimental results show that the proposed multi-objective approach performs consistently better than existing ones.\n\nAuthor summaryThe effects of a treatment are often not the same for different individuals with different gene expressions. Learning to predict the heterogeneous treatment effects from clinical and expression data is an important step towards personalized medical treatment. Existing computational methods are not ideal for the task because they do not address the interpretability of the model and do not consider the limited sample sizes in biological and medical applications. Our method addresses these issues and achieves superior performance in analyzing the treatment effects of radiotherapy on breast cancer patients.

bioinformatics↗

bioRxiv2018

Brain age prediction of healthy subjects on anatomic MRI with deep learning: going beyond with an \"explainable AI\" mindset

ObjectivesDefine a clinically usable preprocessing pipeline for MRI data\n\nPredict brain age using various machine learning and deep learning algorithms\n\nDefine Caveat against common machine learning traps\n\nData and MethodsWe used 1597 open-access T1 weighted MRI from 24 hospitals.\n\nPreprocessing consisted in applying : N4 bias field correction, registration to MNI152 space, white and grey stripe intensity normalization, skull stripping and brain tissue segmentation\n\nPrediction of brain age was done with growing complexity of data input (histograms, grey matter from segmented MRI, raw data) and models for training (linear models, non linear model such as gradient boosting over decision trees, and 2D and 3D convolutional neural networks).\n\nWork on interpretability consisted in (i) proceeding on basic data visualization like correlations maps between age and voxels value, and generating (ii) weights maps of simpler models, (iii) heatmaps from CNNs model with occlusion method.\n\nResultsProcessing time seemed feasible in a radiological workflow : 5 min for one 3D T1 MRI.\n\nWe found a significant correlation between age and gray matter volume with a correlation r = -0.74. Our best model obtained a mean absolute error of 3.60 years, with fine tuned convolution neural network (CNN) pretrained on ImageNet.\n\nWe carefully analyzed and interpreted the center effect.\n\nOur work on interpretability on simpler models permitted to observe heterogeneity of prediction depending on brain regions known for being involved in ageing (grey matter, ventricles). Occlusion method of CNN showed the importance of Insula and deep grey matter (thalami, caudate nuclei) in predictions.\n\nConclusionsPredicting the brain age using deep learning could be a standardized metric usable in daily neuroradiological reports. An explainable algorithm gives more confidence and acceptability for its use in practice. More clinical studies using this new quantitative biomarker in neurological diseases will show how to use it at its best.\n\nFOREWORDO_ST_ABSAbout OwkinC_ST_ABSOWKIN was co-founded in 2016 by Thomas Clozel, MD, a clinical research doctor and former assistant professor in clinical hematology and Gilles Wainrib, PhD, a pioneer in the field of Artificial Intelligence in biology. OWKIN passed the proof-of-concept phase and is now providing its innovative AI algorithms to several of the largest cancer centers and pharmaceutical companies in Europe and in the US. With offices in New York and Paris, we pride ourselves in building a company culture around transparency, collaboration, challenge, optimism and fun.\n\nOwkins teamOwkins team is international, multidisciplinary with incredible talent in machine learning, medicine and business. Our data scientists are among the best in the world, with several Kaggle Masters (top global 100), a DREAM Challenge top performer, and publications in ICML, NIPS and other top scientific journals.\n\nTasks repartitionIdea : Thomas Clozel, Roger Stup, Simon Jegou, Paul Herent\n\nBibliography : Paul Herent, Simon Jegou, Thomas Clozel\n\nData access : Simon Jegou, Paul Herent\n\nData cleaning : Simon Jegou, Paul Herent\n\nData analysis : Simon Jegou, Paul Herent\n\nData preprocessing : Simon Jegou, Paul Herent\n\nData analysis : Simon Jegou, Paul Herent\n\nTraining of models : Simon Jegou, Paul Herent\n\nWork on interpretability : Simon Jegou, Paul Herent\n\nWriting : Paul Herent\n\nRereading : Simon Jegou, Thomas Clozel, Julien Savatovsky, Roger Stupp, Olivier Elemento, Kim Gillier\n\nSubmission to medical congress : Paul Herent, Simon Jegou\n\nThanks to...\n\nSimon Jegou, for your mentoring in Machine Learning,\n\nThomas Clozel and Gilles Wainrib, for your welcome at Owkin, very benevolent, Roger Stupp, for your support and re-reading,\n\nJulien Savatovsky, for your support and re-reading,\n\nValentin Ame and Sylvain Toldo, for your help on the beautiful figures and the design of the related blogpost,\n\nAll the Owkin team members, for the great team work we did (and hope well do) between Paris and New York : Anna Huyghues Despointes, Anna I. Bondarenko, Pierre Courtiol, Derek T. Russell-Kraft, Cedric Whitney, Meriem Sefta, Vincent Lepage, Adrian Gonzalez, Maxime HE,Paul Jehanno, Raphael Leger, Alicia Simion, Eric Tramel, Mikhail Zaslavskiy, Pierre Manceron, Chloe Simpson, Paul Mabillot, Valentin Ame, Mathieu Galtier, Camille Marini, Sylvain Toldo, Charlie Saillard, Olivier Dehaene, Olivier Moindrot,\n\nPascal Roux, for your support, your help, your advices,\n\nAxelle, for your patience and support.

neuroscience↗

bioRxiv2017

The impact of tumor receptor heterogeneity on the response to anti-angiogenic cancer treatment

Multiple promoters and inhibitors mediate angiogenesis, the formation of new blood vessels, and these factors represent potential targets for impeding vessel growth in tumors. Vascular endothelial growth factor (VEGF) is a potent angiogenic factor targeted in anti-angiogenic cancer therapies. In addition, thrombospondin-1 (TSP1) is a major endogenous inhibitor of angiogenesis, and TSP1 mimetics are being developed as an alternative type of anti-angiogenic agent. The combination of bevacizumab, an anti-VEGF agent, and ABT-510, a TSP1 mimetic, has been tested in clinical trials to treat advanced solid tumors. However, the patients responses are highly variable and show disappointing outcomes. To obtain mechanistic insight into the effects of this combination anti-angiogenic therapy, we have constructed a novel whole-body systems biology model including the VEGF and TSP1 reaction networks. Using this molecular-detailed model, we investigated how the combination anti-angiogenic therapy changes the amounts of pro-angiogenic and anti-angiogenic complexes in cancer patients. We particularly focus on answering the question of how the effect of the combination therapy is influenced by tumor receptor expression, one aspect of patient-to-patient variability. Overall, this model complements the clinical administration of combination anti-angiogenic therapy, highlights the role of tumor receptor variability in the heterogeneous responses to anti-angiogenic therapy, and identifies the tumor receptor profiles that correlate with a high likelihood of a positive response to the combination therapy. Our model provides novel understanding of the VEGF-TSP1 balance in cancer patients at the systems-level and could be further used to optimize combination anti-angiogenic therapy.

systems biology↗