Philippe Veber :: research
research | teaching | programming | linux
I am a CNRS research engineer working at the Laboratoire de Biométrie et Biologie Évolutive (LBBE) in Lyon, France. I also participate to the service activity of the PRABI bioinformatics platform. My research interests are on the construction of genome-wide genetic regulatory networks, and especially on how (non-coding) genomic sequences direct transcriptional responses. I am therefore particularly involved in next-gen sequencing techniques as a means to gather sufficient data for developing predictive models of transcription.
Next-gen sequencing analysis
Deep-sequencing is used to provide genome-wide information on transcription: ChIP-seq is used for detecting transcription factor binding and epigenetic modifications, RNA-seq for the identification and quantification of mRNAs. I have developped an ocaml library named Guizmin, which can be used to build pipelines for this kind of data. It can be used directly within an ocaml program as a highlevel abstraction for connecting pipeline components.
Collaborations:
- Gérard Benoit, Centre de Génétique et de Physiologie Moléculaire et Cellulaire
- Sabine Colnot, Institut Cochin, Paris
- Florence d'Alché, Université d'Evry
- Nicolas Brunel, Université d'Evry
Up
Sign Consistency on Influence Graphs
This is a framework that I developed during my thesis, and which can be used to combine genome-wide expression profiles with a qualitative description of genetic regulations, in order to test if they are compatible in some sense. If not, we developed several techniques for diagnosis and model repair.
An implementation based on Answer Set Programming is available as part of the BioASP tools.
Collaborations:
- Torsten Schaub, Potsdam Universität
- Anne Siegel, IRISA, Rennes
Up
Information Extraction for molecular biology
The goal of information extraction (IE) is to produce formal representations from unstructured documents like articles or speech recording. I worked on the extraction of biomolecular events or statements from scientific publications in microbiology: DNA binding, regulations, phenotype of mutants, etc … Our approach is based on the combination of advanced NLP tools and machine learning techniques, like support vector machines and kernel methods.
- this research activity is described on these slides.