Emmanuel Barillot

Genomic Profiling and Identification of High-Risk Uveal Melanoma by Array CGH Analysis of Primary Tumors and Liver Metastases

PURPOSE Incurable metastases develop in approximately 50% of patients with uveal melanoma (UM). The purpose of this study was to analyze genomic profiles in a large series of ocular tumors and liver metastases and design a genome-based classifier for metastatic risk assessment. METHODS A series of 86 UM tumors and 66 liver metastases were analyzed by using a BAC CGH (comparative genomic hybridization) microarray. A clustering was performed, and correlation with the metastatic status was sought among a subset of 71 patients with a minimum follow-up of 24 months. The status of chromosome 3 was further examined in the tumors, and metastases with disomy 3 were checked with an SNP microarray. A prognostic classifier was constructed using a log-linear model on minimal regions and leave-one-out cross-validation. RESULTS The clustering divides the groups of tumors with disomy 3 and monosomy 3 into two and three subgroups, respectively. Same subgroups are found in primary tumors and in metastases, but with different frequencies. Isolated monosomy 3 was present in 0% of metastatic ocular tumors and in 3% of metastases. The highest metastatic rate in ocular tumors was observed in a subgroup defined by the gain of 8q with a proximal breakpoint, and losses of 3, 8p, and 16q, also most represented in metastases. A prognostic classifier that included the status of these markers led to an 85.9% classification accuracy. CONCLUSIONS The analysis of the status of these specific chromosome regions by genome profiling on SNP microarrays should be a reliable tool for identifying high-risk patients in future adjuvant therapy protocols.

Advanced spot quality analysis in two-colour microarray experiments

BackgroundImage analysis of microarrays and, in particular, spot quantification and spot quality control, is one of the most important steps in statistical analysis of microarray data. Recent methods of spot quality control are still in early age of development, often leading to underestimation of true positive microarray features and, consequently, to loss of important biological information. Therefore, improving and standardizing the statistical approaches of spot quality control are essential to facilitate the overall analysis of microarray data and subsequent extraction of biological information.FindingsWe evaluated the performance of two image analysis packages MAIA and GenePix (GP) using two complementary experimental approaches with a focus on the statistical analysis of spot quality factors. First, we developed control microarrays with a priori known fluorescence ratios to verify the accuracy and precision of the ratio estimation of signal intensities. Next, we developed advanced semi-automatic protocols of spot quality evaluation in MAIA and GP and compared their performance with available facilities of spot quantitative filtering in GP. We evaluated these algorithms for standardised spot quality analysis in a whole-genome microarray experiment assessing well-characterised transcriptional modifications induced by the transcription regulator SNAI1. Using a set of RT-PCR or qRT-PCR validated microarray data, we found that the semi-automatic protocol of spot quality control we developed with MAIA allowed recovering approximately 13% more spots and 38% more differentially expressed genes (at FDR = 5%) than GP with default spot filtering conditions.ConclusionCareful control of spot quality characteristics with advanced spot quality evaluation can significantly increase the amount of confident and accurate data resulting in more meaningful biological conclusions.

The inconvenience of data of convenience: Computational research beyond post-mortem analyses

The inconvenience of data of convenience: computational research beyond post-mortem analyses

Comparative analyses of super-enhancers reveal conserved elements in vertebrate genomes

Super-enhancers (SEs) are key transcriptional drivers of cellular, developmental, and disease states in mammals, yet the conservational and regulatory features of these enhancer elements in nonmammalian vertebrates are unknown. To define SEs in zebrafish and enable sequence and functional comparisons to mouse and human SEs, we used genome-wide histone H3 lysine 27 acetylation (H3K27ac) occupancy as a primary SE delineator. Our study determined the set of SEs in pluripotent state cells and adult zebrafish tissues and revealed both similarities and differences between zebrafish and mammalian SEs. Although the total number of SEs was proportional to the genome size, the genomic distribution of zebrafish SEs differed from that of the mammalian SEs. Despite the evolutionary distance separating zebrafish and mammals and the low overall SE sequence conservation, ∼42% of zebrafish SEs were located in close proximity to orthologs that also were associated with SEs in mouse and human. Compared to their nonassociated counterparts, higher sequence conservation was revealed for those SEs that have maintained orthologous gene associations. Functional dissection of two of these SEs identified conserved sequence elements and tissue-specific expression patterns, while chromatin accessibility analyses predicted transcription factors governing the function of pluripotent state zebrafish SEs. Our zebrafish annotations and comparative studies show the extent of SE usage and their conservation across vertebrates, permitting future gene regulatory studies in several tissues.

PhysiBoSS: a multi-scale agent-based modelling framework integrating physical dimension and cell signalling

Abstract Motivation Due to the complexity and heterogeneity of multicellular biological systems, mathematical models that take into account cell signalling, cell population behaviour and the extracellular environment are particularly helpful. We present PhysiBoSS, an open source software which combines intracellular signalling using Boolean modelling (MaBoSS) and multicellular behaviour using agent-based modelling (PhysiCell). Results PhysiBoSS provides a flexible and computationally efficient framework to explore the effect of environmental and genetic alterations of individual cells at the population level, bridging the critical gap from single-cell genotype to single-cell phenotype and emergent multicellular behaviour. PhysiBoSS thus becomes very useful when studying heterogeneous population response to treatment, mutation effects, different modes of invasion or isomorphic morphogenesis events. To concretely illustrate a potential use of PhysiBoSS, we studied heterogeneous cell fate decisions in response to TNF treatment. We explored the effect of different treatments and the behaviour of several resistant mutants. We highlighted the importance of spatial information on the population dynamics by considering the effect of competition for resources like oxygen. Availability and implementation PhysiBoSS is freely available on GitHub (https://github.com/sysbio-curie/PhysiBoSS), with a Docker image (https://hub.docker.com/r/gletort/physiboss/). It is distributed as open source under the BSD 3-clause license. Supplementary information Supplementary data are available at Bioinformatics online.

Clonal assessment of functional mutations in cancer based on a genotype-aware method for clonal reconstruction

In cancer, clonal evolution is characterized based on single nucleotide variants and copy number alterations. Nonetheless, previous methods failed to combine information from both sources to accurately reconstruct clonal populations in a given tumor sample or in a set of tumor samples coming from the same patient. Moreover, previous methods accepted as input all variants predicted by variant-callers, regardless of differences in dispersion of variant allele frequencies (VAFs) due to uneven depth of coverage and possible presence of strand bias, prohibiting accurate inference of clonal architecture. We present a general framework for assignment of functional mutations to specific cancer clones, which is based on distinction between passenger variants with expected low dispersion of VAF versus putative functional variants, which may not be used for the reconstruction of cancer clonal architecture but can be assigned to inferred clones at the final stage. The key element of our framework is QuantumClone, a method to cluster variants into clones, which we have thoroughly tested on simulated data. QuantumClone takes into account VAFs and genotypes of corresponding regions together with information about normal cell contamination. We applied our framework to whole genome sequencing data for 19 neuroblastoma trios each including constitutional, diagnosis and relapse samples. We discovered specific pathways recurrently altered by deleterious mutations in different clonal populations. Some such pathways were previously reported (e.g., MAPK and neuritogenesis) while some were novel (e.g., epithelial-mesenchymal transition, cell survival and DNA repair). Most pathways and their modules had more mutations at relapse compared to diagnosis.

Signalling maps in cancer research: construction and data analysis

Abstract Generation and usage of high-quality molecular signalling network maps can be augmented by standardizing 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.

Effective normalization for copy number variation in Hi-C data

BackgroundNormalization is essential to ensure accurate analysis and proper interpretation of sequencing data, and chromosome conformation capture data such as Hi-C have particular challenges. Although several methods have been proposed, the most widely used type of normalization of Hi-C data usually casts estimation of unwanted effects as a matrix balancing problem, relying on the assumption that all genomic regions interact equally with each other.ResultsIn order to explore the effect of copy-number variations on Hi-C data normalization, we first propose a simulation model that predict the effects of large copy-number changes on a diploid Hi-C contact map. We then show that the standard approaches relying on equal visibility fail to correct for unwanted effects in the presence of copy-number variations. We thus propose a simple extension to matrix balancing methods that model these effects. Our approach can either retain the copy-number variation effects (LOIC) or remove them (CAIC). We show that this leads to better downstream analysis of the three-dimensional organization of rearranged genomes.ConclusionsTaken together, our results highlight the importance of using dedicated methods for the analysis of Hi-C cancer data. Both CAIC and LOIC methods perform well on simulated and real Hi-C data sets, each fulfilling different needs.

Biological network modelling and precision medicine in oncology

Precision medicine in oncology is becoming reality thanks to the next-generation sequencing of tumours and the development of targeted inhibitors enabling tailored therapies. Many clinical trials base their strategy on the identification of mutations to deliver the targeted inhibitor that counteract supposedly the effect of a mutated gene. Recent results have shown that this gene-centered strategy can be successful, but can also fall short in stopping progression. This is due to the many compensation mechanisms, cross-talks and feedback loops that enable the tumoral cell to escape treatment. Taking into account the regulatory network is necessary to establish which inhibitor or combination of inhibitors would achieve the best therapeutic results. Mathematical modelling of biological networks, together with high-quality pathway databases collecting our knowledge of the molecular circuitry of normal and tumoral cells, hold the hopes of an enhanced future for precision medicine in oncology.

How Cell Decides Between Life and Death: Mathematical Modeling of Epigenetic Landscapes of Cellular Fates

We present a mathematical model of cell fate decision between survival, necrosis and apoptosis as a concrete implementation of the Waddington’s metaphor of epigenetic landscape determining cellular behaviour. We describe the principles of the model construction and in silico experiments performed on it. The genetic network underlying cell fate decisions is reconstructed in the form of an influence diagram together with logical rules defining possible system state changes, while the epigenetic landscape is represented as a state transition graph generated from the discrete cell fate decision model. Stochastic cellular decision making is modeled as a random walk on the state transition graph, assuming equal probabilities of any possible system state update. The probability to reach a particular attractor (stable state) of the state transition graph, starting the random walk from a “physiological” cellular state, is interpreted as a probability of having a particular phenotype (outcome) in a biological experiment. As a result, one can predict the phenotypic probabilities and their changes as a result of specific perturbations. We show that such a “generic” cell model can recapitulate and explain experiments conducted on mice and cell lines and predict the outcome of not yet done experiments. In the discussion, we compare the design principles of the cell fate decision model with the principles of designing engineered devices and underline some important differences.