International audienceBig Data is one of the major challenges of statistical science and has numerous consequences from algorithmic and theoretical viewpoints. Big Data always involve massive data but they also often include online data and data heterogeneity. Recently some statistical methods have been adapted to process Big Data, like linear regression models, clustering methods and bootstrapping schemes. Based on decision trees combined with aggregation and bootstrap ideas, random forests were introduced by Breiman in 2001. They are a powerful nonparametric statistical method allowing to consider in a single and versatile framework regression problems, as well as two-class and multi-class classification problems. Focusing on classification problems, this paper proposes a selective review of available proposals that deal with scaling random forests to Big Data problems. These proposals rely on parallel environments or on online adaptations of random forests. We also describe how related quantities -- such as out-of-bag error and variable importance -- are addressed in these methods. Then, we formulate various remarks for random forests in the Big Data context. Finally, we experiment five variants on two massive datasets (15 and 120 millions of observations), a simulated one as well as real world data. One variant relies on subsampling while three others are related to parallel implementations of random forests and involve either various adaptations of bootstrap to Big Data or to ``divide-and-conquer'' approaches. The fifth variant relates on online learning of random forests. These numerical experiments lead to highlight the relative performance of the different variants, as well as some of their limitations

Motivation: Recent high-throughput sequencing advances have expanded the breadth of available omics datasets and the integrated analysis of multiple datasets obtained on the same samples has allowed to gain important insights in a wide range of applications. However, the integration of various sources of information remains a challenge for systems biology since produced datasets are often of heterogeneous types, with the need of developing generic methods to take their different specificities into account. Results: We propose a multiple kernel framework that allows to integrate multiple datasets of various types into a single exploratory analysis. Several solutions are provided to learn either a consensus metakernel or a meta-kernel that preserves the original topology of the datasets. We applied our framework to analyse two public multi-omics datasets. First, the multiple metagenomic datasets, collected during the TARA Oceans expedition, was explored to demonstrate that our method is able to retrieve previous findings in a single KPCA as well as to provide a new image of the sample structures when a larger number of datasets are included in the analysis. To perform this analysis, a generic procedure is also proposed to improve the interpretability of the kernel PCA in regards with the original data. Second, the multi-omics breast cancer datasets, provided by The Cancer Genome Atlas, is analysed using a kernel Self-Organizing Maps with both single and multi-omics strategies. The comparison of this two approaches demonstrates the benefit of our integration method to improve the representation of the studied biological system. Contact: jerome.mariette@inra.fr and nathalie.villa-vialaneix@inra.fr Availability and implementation: Proposed methods are available in the R package mixKernel, released on CRAN. It is fully compatible with the mixOmics package and a tutorial describing the approach can be found on mixOmics web site http://mixomics.org/mixkernel/.

The environmental costs of intensive farming activities are often underestimated or not traded by the market, even though they play an important role in addressing future society's needs. The estimation of nitrogen (N) dynamics is thus an important issue which demands detailed simulation based methods and their integrated use to correctly represent complex and nonlinear interactions into cropping systems. To calculate the N 2 O flux and N leaching from European arable lands, a modeling framework has been developed by linking the CAPRI agro-economic dataset with the DNDC-EUROPE bio-geo-chemical model. But, despite the great power of modern calculators, their use at continental scale is often too computationally costly. By comparing several statistical methods this paper aims to design a metamodel able to approximate the expensive code of the detailed modeling approach, devising the best compromise between estimation performance and simulation speed. We describe the use of two parametric (linear) models and six nonparametric approaches: two methods based on splines (ACOSSO and SDR), one method based on kriging (DACE), a neural networks method (multilayer perceptron, MLP), SVM and a bagging method (random forest, RF). This analysis shows that, as long as few data are available to train the model, splines approaches lead to best results, while when the size of training dataset increases, SVM and RF provide faster and more accurate solutions.