My work lies at the frontier of statistical machine learning, information theory and biologyThe ultimate goal of such research is to identify the right genes to target/alter in view of improving our lives with living cells (such as bacteria, fungi, algae and plants).
More precisely, my research focuses on reverse engineering of cells and genomes using concepts such as causality, implication, redundancy and complementarity (or synergy) of variables in network inference and feature selection algorithms

Main Scientific Contributions
  • Network Inference and Meta-analysis from expression data
    • I invented the MRNet Network inference method (2007) under G. Bontempi supervision, one of the first method i) able to infer a large-scale causal network but also ii) that uses a feature selection algorithm in order to build a network (a strategy that has now become quite common). I also produced a variant called MRNetB (2010), together with MIT professor M. Kellis.  
    • I supervised my two first PhD students P. Bellot and N. C. Pham, to study and produce meta-analysis methods for network inference. Our various papers have been selected by well known editors to become book chapters (2017-2018) and a "best paper award" got re-published in the BMC BioDatamining journal (2017). 
  • Systems Biology and Computational Biology
    • I produced the first comprehensive Drosophila transcriptional network, allowing me to become a co-first author of the AAAS Science modENCODE paper (2010) and its related Genome Research paper (2012). Beside its unusual large scale size for the time (i.e. 200k nodes), my transcriptional network has been the very first to integrate chromatin data into its inference. Those two papers keep producing more than 100 citations per year. 
    • Together with G. Lobet, my PhD student M. Noll produced the first automatic, machine-learning-based plant-root annotation system. This contribution lead us to two journal publications: one in Frontiers in Plant Science (2017) and one in GIGAscience (2017).
  • Open-Source Bioinformatics Tools
    • I have authored the MINet Bioconductor package (2008), one of the two first R package able to infer large-scale (up to several thousands) networks from Data. Ten years later, the package reaches between 5000 and 10000 downloads a year and its related paper still get the unusual rate of 30 new citations per year. 
    • I have also authored the infotheo R package (2009), this package allows to compute information-theoretic measure in R and is among the top downloaded package of the CRAN ecosystem. 
    • My PhD student P. Bellot and I produced the Netbenchmark Bioconductor package (2015), a tool that allows a thorough and fast comparison of network inference algorithms. Two years later its related paper is in the top 25% (of all research outputs scored by Altmetric) for papers of the same age and source.
  • Feature Selection and machine learning
    • From 2005 to 2008, I authored several peer-reviewed papers on feature selection. Together with G. Bontempi, we invented the two feature selection methods DISR (2006) and, with C. Schretter, its generic formulation called MASSIVE (2008).  Those two information-theoretic filters have been among the very first in the field to use explicitly the concept of complementarity (or synergy) of variables. This concept of complementarity (or synergy) of variables, almost forgotten during 50 years, have since been widely used in feature selection leading our related papers to reach the key threshold of 100 citations.
  • Causality and Implications in Data Analysis
    • I co-authored with G. Bontempi an ICML paper introducing the mIMR feature selection method (2010), one of the first method to focus on selecting causal variables (instead of just informative ones).
    • I have proposed the Rank Minrelation Coefficient (2014), the first statistical measure focused on selecting implicative variables (from continuous data). Since then, I have given several talks on this theoretical topic of causality and implication, and to this day, my lab keeps exploring the intricate connection between these two key concepts.
  • Robotics and automation
    • Under Marco Dorigo supervision, I produced a study of collective retrieval of an heavy object by very "simple" autonomous robots (2004). Our lab now develop Arduino-based and raspberry-pi-based system of regulation and automation for growing plants, algae and fishes.
Selected Publications
  1. NetBenchmark: A Bioconductor Package for Reproducible Benchmarks of Gene Regulatory Network Inference.
    Pau Bellot, Catharina Olsen, Philippe Salembier, Albert Oliveras-Verges and Patrick E. Meyer.
    BMC Bioinformatics 2015.
  2. The Rank Minrelation Coefficient.
    Patrick E. Meyer.
    Quality Technology & Quantitative Management Vol. 11. 2014. 
    Special Issue on "Graphical Causality Models: Trees, Bayesian Networks and Big Data" (ENBIS-SFdS Meeting)

  3. Identification of Functional Elements and Regulatory Circuits by Drosophila modENCODE.
    The modENCODE Consortium et al. (co-first author)
    In Science (AAAS), 2010.
    Press: Science Daily - Le Figaro.
    preversion [pdf]

  4. MINET: An open source R/Bioconductor Package for Mutual Information based Network Inference.
    Patrick E. Meyer, Frederic Lafitte and Gianluca Bontempi.
    In BMC Bioinformatics, Volume 9, 2008.

  5. Information-Theoretic Feature Selection in Microarray Data Using Variable Complementarity.
    Corrected version [pdf]
    Patrick E. Meyer, Colas Schretter and Gianluca Bontempi.
    In IEEE Journal of Selected Topics in Signal Processing (JSTSP)
    Special Issue on Genomic and Proteomic Signal Processing, Volume2, Issue 3, 2008

Other Peer-Reviewed Publications

  1. Gene Entity Recognition of Full Text Articles
    M Noll, J Lété, P Meyer
    ICBBS'17 Proceedings of the 6th International Conference on Bioinformatics and Biomedical Sciences, 2017
  2. Combining semi-automated image analysis techniques with machine learning algorithms to accelerate large scale genetic studies.
    JA Atkinson, G Lobet, M Noll, PE Meyer, M Griffiths, DM Wells
    GigaScience, 2017
  3. Using a Structural Root System Model to Evaluate and Improve the Accuracy of Root Image Analysis Pipelines
    Guillaume Lobet, Iko T. Koevoes, Manuel Noll, Patrick E. Meyer, Pierre Tocquin, Loïc Pagès, and Claire Périlleux.
    Frontiers in Plant Science, 2017
  4. Meta-Analysis Strategies for Information-Theoretic Network Inference.
    C. Pham, Benjamin Haibe-Kains, Pau Bellot, Gianluca Bontempi and Patrick E. Meyer.
    In BMC BioData Mining, 2017
    Best paper award of the IEEE DEXA Workshops 2016. preversion [pdf]

  5. Study of normalization and aggregation approaches for consensus network estimation.
    Pau Bellot, Philippe Salembier, Albert Oliveras-Verges and Patrick E Meyer.
    In Proceedings of IEEE Alife 2015.preversion [pdf]

  6. Efficient combination of pairwise feature networks.
    Pau Bellot and Patrick E. Meyer, in NCW ECML2014 (JMLR workshop), Nancy, France 2014. 
    preversion [pdf]

  7. Predictive regulatory models in Drosophila melanogaster by integrative inference of transcriptional networks.
    Daniel Marbach, Sushmita Roy, Ferhat Ay, Patrick E. Meyer, Rogerio Candeias, Tamer Kahveci, Christopher A. Bristow and Manolis Kellis.
    In Genome Research, 2012. 
    preversion [pdf]

  8. Estimation of temporal lags for the inference of gene regulatory networks from time series.
    Miguel Lopes, Patrick E. Meyer and Gianluca Bontempi.
    In BENELEARN'12, 2012.

  9. Causal Filter Selection in Microarray Data.
    Gianluca Bontempi and Patrick E. Meyer
    In ICML’10, International Conference On Machine Learning, 2010.
    preversion [pdf]

  10. Information-Theoretic Inference of Gene Networks Using Backward Elimination.
    Patrick E. Meyer, Daniel Marbach, Sushmita Roy and Manolis Kellis.
    In BioComp'10, International Conference on Bioinformatics and Computational Biology, 2010.
    preversion [pdf]

  11. On the Impact of Entropy Estimation on Transcriptional Regulatory Network Inference Based on Mutual Information
    Catharina Olsen, Patrick E. Meyer and Gianluca Bontempi.
    In EURASIP Journal on Bioinformatics and Systems Biology, 2009

  12. Information-Theoretic Inference of Large Transcriptional Regulatory Networks
    Patrick E. Meyer, Kevin Kontos, Frederic Lafitte and Gianluca Bontempi.
    In EURASIP Journal on Bioinformatics and Systems Biology
    Special Issue on Information-Theoretic Methods for Bioinformatics, 2007 (highly accessed)

  13. A Model-Based Relevance Estimation Approach for Feature Selection in Microarray Datasets
    Gianluca Bontempi and Patrick E. Meyer.
    In ICANN'08, International Conference on Artificial Neural Networks, 2008.
    Lecture Notes in Computer Science, volume 5164, pp. 21-31, Springer, 2008

  14. Biological Network Inference Using Redundancy Analysis.
    Patrick E. Meyer, Kevin Kontos and Gianluca Bontempi.
    In BIRD'07, 1st International Conference on Bioinformatics Research and Development, 2007.
    Lecture Notes in Computer Science, volume 4414, pp. 16-27, Springer, 2007
    S. Hochreiter and R. Wagner editors.
    preversion [pdf] 

  15. On the Use of Variable Complementarity for Feature Selection in Cancer Classification.
    Patrick E. Meyer and Gianluca Bontempi.
    In EvoBIO'06, 4th European Conference on Evolutionary Computation, Machine Learning and Data Mining in Bioinformatics, 2006.
    Lecture Notes in Computer Science, volume 3907, pp. 91-102, Springer, 2006
    F. Rothlauf et al. editors.
    preversion [pdf] 

  16. Combining Lazy Learning, Racing and Subsampling for Effective Feature Selection
    Gianluca Bontempi, Mauro Birattari and Patrick E. Meyer.
    In ICANNGA'05, 7th International Conference on Adaptive and Natural Computing Algorithms, 2005.
    Springer, pp. 393-396, 2005.
    B. Ribeiro et al. editors

  17. Speeding up Feature Selection by Using an Information Theoretic Bound
    Patrick E. Meyer, Olivier Caelen and Gianluca Bontempi.
    In BNAIC'05, 17th Belgian-Dutch Conference on Artificial Intelligence, 2005.
    KVAB, pp.166-173, 2005.
    preversion [pdf] 

  18. Collective Retrieval by Autonomous Robots
    Patrick E. Meyer.
    In STAIRS-ECAI'04, 2th Starting AI Researchers' Symposium, 2004.
    Frontiers in Artificial Intelligence and Applications, volume 109, pp. 199-204, IOS Press, 2004
    E. Onaindia and S. Staab editors.
    preversion [pdf]

Book Chapters

  1. Unsupervised GRN Ensemble.
    Pau Bellot, Philippe Salembier, Ngoc C. Pham and Patrick E. Meyer,
    Gene Regulatory Networks
    Guido Sanguinetti and Vân Anh Huynh-Thu Eds. Springer 2018.

  2. Efficient combination of pairwise feature networks.
    Pau Bellot and Patrick E. Meyer,
    Neural Connectomics Challenge.
    Battaglia et al. Eds. Springer 2017.
  3. Information-Theoretic Gene Selection in Expression Data
    Patrick E. Meyer and Gianluca Bontempi.
    Biological Knowledge Discovery Handbook: Preprocessing, Mining and Postprocessing of Biological Data.
    Elloumi et al. Editors. Wiley, 2013.
    preversion [pdf]

  4. Transcriptional Network Inference based on Information Theory.
    Patrick E. Meyer, Catharina Olsen and Gianluca Bontempi.
    Applied Statistics for Network Biology: Methods in Systems Biology. 
    Dehmer et al. Editors. Wiley, 2011. ISBN: 978-3-527-32750-8. 
    preversion [pdf]

Conferences - Workshops - Fact sheets

  • Meta-Analysis of Transcriptional Network Inference
    Patrick E. Meyer, Benjamin Haibe-Kains and Gianluca Bontempi
    Recomb Satellite 09, MIT.

  • Fact sheet: Using mutual information to infer causal relationships
    Catharina Olsen, Patrick E. Meyer and Gianluca Bontempi
    JMLR: Workshop and Conference Proceedings - NIPS 2008 workshop on causality.
    file [pdf]

Theses and Books

  • Information-Theoretic Variable Selection and Network Inference from Microarray Data
    Ph.D. Thesis (Doctorat) in Sciences, Machine Learning Group, Universite Libre de Bruxelles, Belgium, 2008.
    file (5Mb) [pdf]