Céline Rouveirol

Computation of recurrent minimal genomic alterations from array-CGH data

MOTIVATION The identification of recurrent genomic alterations can provide insight into the initiation and progression of genetic diseases, such as cancer. Array-CGH can identify chromosomal regions that have been gained or lost, with a resolution of approximately 1 mb, for the cutting-edge techniques. The extraction of discrete profiles from raw array-CGH data has been studied extensively, but subsequent steps in the analysis require flexible, efficient algorithms, particularly if the number of available profiles exceeds a few tens or the number of array probes exceeds a few thousands. RESULTS We propose two algorithms for computing minimal and minimal constrained regions of gain and loss from discretized CGH profiles. The second of these algorithms can handle additional constraints describing relevant regions of copy number change. We have validated these algorithms on two public array-CGH datasets. AVAILABILITY From the authors, upon request. CONTACT celine@lri.fr SUPPLEMENTARY INFORMATION Supplementary data are available at Bioinformatics online.

Supervised Machine Learning Applied to Link Prediction in Bipartite Social Networks

This work copes with the problem of link prediction in large-scale two-mode social networks. Two variations of the link prediction tasks are studied: predicting links in a bipartite graph and predicting links in a unimodal graph obtained by the projection of a bipartite graph over one of its node sets. For both tasks, we show in an empirical way, that taking into account the bipartite nature of the graph can enhance substantially the performances of prediction models we learn. This is achieved by introducing new variations of topological atttributes to measure the likelihood of two nodes to be connected. Our approach, for both tasks, consists in expressing the link prediction problem as a two class discrimination problem. Classical supervised machine learning approaches can then be applied in order to learn prediction models. Experimental validation of the proposed approach is carried out on two real data sets: a co-authoring network extracted from the DBLP bibliographical database and bipartite graph history of transactions on an on-line music e-commerce site.

Flattening and Saturation: Two Representation Changes for Generalization

Two representation changes are presented: the first one, called flattening, transforms a first-order logic program with function symbols into an equivalent logic program without function symbols; the second one, called saturation, completes an example description with relevant information with respect to both the example and available background knowledge. The properties of these two represenlation changes are analyzed as well as their influence on a generalization algorithm that takes a single example as input.

Towards Learning in CARIN-ALN

In this paper we investigate a new language for learning, which combines two well-known representation formalisms, Description Logics and Horn Clause Logics. Our goal is to study the feasability of learning in such a hybrid description - horn clause language, namely CARIN-ALN [LR98b], in the presence of hybrid background knowledge, including a Horn clause and a terminological component. After setting our learning framework, we present algorithms for testing example coverage and subsumption between two hypotheses, based on the existential entailment algorithm studied in[LR98b]. While the hybrid language is more expressive than horn clause logics alone, the complexity of these two steps for CARIN-ALN remains bounded by their respective complexity in horn clause logics.

Resource-bounded Relational Reasoning: Induction and Deduction Through Stochastic Matching

One of the obstacles to widely using first-order logic languages is the fact that relational inference is intractable in the worst case. This paper presents an any-time relational inference algorithm: it proceeds by stochastically sampling the inference search space, after this space has been judiciously restricted using strongly-typed logic-like declarations.We present a relational learner producing programs geared to stochastic inference, named STILL, to enforce the potentialities of this framework. STILL handles examples described as definite or constrained clauses, and uses sampling-based heuristics again to achieve any-time learning.Controlling both the construction and the exploitation of logic programs yields robust relational reasoning, where deductive biases are compensated for by inductive biases, and vice versa.

LICORN: learning cooperative regulation networks from gene expression data

MOTIVATION One of the most challenging tasks in the post-genomic era is the reconstruction of transcriptional regulation networks. The goal is to identify, for each gene expressed in a particular cellular context, the regulators affecting its transcription, and the co-ordination of several regulators in specific types of regulation. DNA microarrays can be used to investigate relationships between regulators and their target genes, through simultaneous observations of their RNA levels. RESULTS We propose a data mining system for inferring transcriptional regulation relationships from RNA expression values. This system is particularly suitable for the detection of cooperative transcriptional regulation. We model regulatory relationships as labelled two-layer gene regulatory networks, and describe a method for the efficient learning of these bipartite networks from discretized expression data sets. We also evaluate the statistical significance of such inferred networks and validate our methods on two public yeast expression data sets. AVAILABILITY http://www.lri.fr/~elati/licorn.html. SUPPLEMENTARY INFORMATION Supplementary data are available at Bioinformatics online.

A supervised machine learning link prediction approach for academic collaboration recommendation

In this work we tackle the problem of link prediction in co-authoring network. We apply a topological dyadic supervised machine learning approach for that purpose. A co-authoring network is actually obtained by the projection of a two-mode graph (an authoring graph linking authors to publications they have signed) over the authors set. We show that link prediction performances can be substantially enhanced by analyzing not only the co-authoring network, but also the dual graph obtained by projecting the original two-mode network over the set of publications.

A case study in a recommender system based on purchase data

Collaborative filtering has been extensively studied in the context of ratings prediction. However, industrial recommender systems often aim at predicting a few items of immediate interest to the user, typically products that (s)he is likely to buy in the near future. In a collaborative filtering setting, the prediction may be based on the users purchase history rather than rating information, which may be unreliable or unavailable. In this paper, we present an experimental evaluation of various collaborative filtering algorithms on a real-world dataset of purchase history from customers in a store of a French home improvement and building supplies chain. These experiments are part of the development of a prototype recommender system for salespeople in the store. We show how different settings for training and applying the models, as well as the introduction of domain knowledge may dramatically influence both the absolute and the relative performances of the different algorithms. To the best of our knowledge, the influence of these parameters on the quality of the predictions of recommender systems has rarely been reported in the literature.

Identification of functional modules based on transcriptional regulation structure

BackgroundIdentifying gene functional modules is an important step towards elucidating gene functions at a global scale. Clustering algorithms mostly rely on co-expression of genes, that is group together genes having similar expression profiles.ResultsWe propose to cluster genes by co-regulation rather than by co-expression. We therefore present an inference algorithm for detecting co-regulated groups from gene expression data and introduce a method to cluster genes given that inferred regulatory structure. Finally, we propose to validate the clustering through a score based on the GO enrichment of the obtained groups of genes.ConclusionWe evaluate the methods on the stress response of S. Cerevisiae data and obtain better scores than clustering obtained directly from gene expression.

Incremental Learning of Relational Action Rules

In the Relational Reinforcement learning framework, we propose an algorithm that learns an action model allowing to predict the resulting state of each action in any given situation. The system incrementally learns a set of first order rules: each time an example contradicting the current model (a counter-example) is encountered, the model is revised to preserve coherence and completeness, by using data-driven generalization and specialization mechanisms. The system is proved to converge by storing counter-examples only, and experiments on RRL benchmarks demonstrate its good performance w.r.t state of the art RRL systems.