Postes PDF Print Email
Written by   
Friday, 05 April 2013 15:32

We are offering a 2 years post-doctoral position, in statistics and evolution

to « improve the comparative framework based on similarity networks”,

starting on March the 1rst 2017.


The EVOLUNET project, funded by the ERC, is enhancing the use of similarity networks, adapted to evolutionary biology questions ( Such networks, inspired by the studies on social and regulatory networks, allow mainly for fast inclusive comparative analyses of both (highly) divergent and conserved molecular sequences from living organisms, mobile genetic elements, that were obtained either from lab cultures or from environmental samples, or both. These networks allow in particular to detect chimeric genes (i.e. genes made of parts with different evolutionary histories) and therefore to investigate the evolution of genes ; and to detect lateral gene transfer, and therefore to investigate the evolution of genomes and microbial communities. Moreover, similarity networks also allow to compare diverse data, gathered by experts from different evolutionary fields, such as historical linguistics and paleontology. The EVOLUNET project already implemented various methods and applied them to different evolutionary questions. Now, network comparison becomes the next important development for network analyses. In particular, network comparison should be most useful to test whether different similarity networks (i.e. constructed from samples from different locations and/or different time periods) have resulted from similar (or from distinct) evolutionary processes.

To further develop network comparisons, especially networks of sequences from environmental microbiomes, our lab is now looking for a post-doctoral fellow in bio-informatics, graph theory or evolutionary biology with a strong background in programming, who will be funded for up to 2 years by an ERC grant (for an approximate salary of 2,727 euros/ month before taxes).

The candidate will take advantage of the properties of many of the real-lie networks constructed in the lab, of our original microbiome data, and of algorithms from network sciences for:

-1) Developing and comparing network-based diversity indices

-2) Constituting a comparative framework between real-life sequence similarity networks and sequence similarity networks generated based on various models of sequence evolution


The first task implies exploring various network-based diversity measures to compare labelled weighted networks that harbor millions of nodes and dozens of millions of edges. Such networks are often too large to be directly compared by graph matching algorithms. However, they contain precious informations regarding genetic diversity of the investigated datasets. While diversity measures have been developed for phylogenetic trees or phylogenetic networks, sequence similarity networks provide an additional kind of distances among biological objects (sequences, genomes, environments) that can be investigated taking advantages of the network topological properties, and therefore the comparison of similarity networks should provide new insights about the evolution of genetic diversity.


The second task relies upon descriptions of large networks by various topological indices (diameter, clustering, distribution of motifs of a given size, …) and by the diversity indices considered above. These descriptions will be used to test whether two networks show the same properties or not. The postdoctoral fellow will first test if real-life similarity networks show the same properties as networks generated from sequences evolved under a given model to determine whether our data fit that particular model. Such comparisons will help the postdoctoral fellow to determine what a real-life network properties usually are, and thus help him/her build better null model similarity networks, specifically designed for evolutionary studies (as opposed to the random networks often used in the graph literature that do not always fit the properties of real-life networks). The postdoctoral fellow will also develop statistics to compare the topological and diversity indices of real-life similarity networks with those of carefully simulated networks. Such statistics will allow classifying genes and proteins into sets that show a diversity that can be relatively simply explained (e.g. displaying indices compatible with known processes and evolutionary models), or require other explanations (e.g. surprisingly extremes rates of divergence, fusion, or recombination). Ultimately, the postdoctoral fellow will determine which simulated networks best fit with real-life networks and develop a random graph model specifically designed for evolutionary studies, so that this model might be further exploited by graph theorists in the field of evolutionary biology.

The candidate will work within a consortium of friendly bioinformaticians (Philippe Lopez, Eduardo Corel), evolutionary biologists (Eric Bapteste), graph theorists (Michel Habib, Laurent Viennot), and statisticians (François-Joseph Lapointe), and be hosted in the University Pierre and Marie Curie in the center of Paris, France.

Ideally, the candidate should have a strong interest for evolutionary biology, microbial evolution and a good background in bio-informatics, or graph theory. The position will start by March 2017, but interested candidates are invited to apply immediately.

He/she should be early in his/her carreer (i.e. holding a PhD degree for less than 2 years).

Applicants are requested to send a detailed resume, a motivation letter, a pdf copy of their PhD thesis, and the names of two scientific referees to :
This e-mail address is being protected from spambots. You need JavaScript enabled to view it

The first round of applications will be closed January 25th, 2017.

Last Updated on Monday, 02 January 2017 13:21