Arnaud Fontaine

Computational identification of protein-coding sequences by comparative analysis

Gene prediction is an essential step in understanding the genome of a species once it has been sequenced. For that, a promising direction in current research on gene finding is a comparative genomics approach. In this paper, we present a novel approach to identifying evolutionarily conserved protein-coding sequences in genomes. The method takes advantage of the specific substitution pattern of coding se- quences together with the consistency of reading frames. It has been implemented in a software called Protea. Large- scale experimentation shows good results. Protea is in- tended to be a useful complement to existing tools based on homology search or statistical properties of the sequences.

MAGNOLIA: multiple alignment of protein–coding and structural RNA sequences

MAGNOLIA is a new software for multiple alignment of nucleic acid sequences, which are recognized to be hard to align. The idea is that the multiple alignment process should be improved by taking into account the putative function of the sequences. In this perspective, MAGNOLIA is especially designed for sequences that are intended to be either protein-coding or structural RNAs. It extracts information from the similarities and differences in the data, and searches for a specific evolutionary pattern between sequences before aligning them. The alignment step then incorporates this information to achieve higher accuracy. The website is available at http://bioinfo.lifl.fr/magnolia.

On-device control flow verification for Java programs

While mobile devices have become ubiquitous and generally multi-application capable, their operating systems provide few high level mechanisms to protect services offered by application vendors against potentially hostile applications coexisting on the device. In this paper, we tackle the issue of controlling application interactions including collusion in Java-based systems running on open, constrained devices such as smart cards or mobile phones. We present a model specially designed to be embedded in constrained devices to verify on-device at loading-time that interactions between applications abide by the security policies of each involved application without resulting in run-time computation overheads; this model deals with application (un)installations and policy changes in an incremental fashion. We sketch the application of our approach and its security enhancements on a multi-application use case for GlobalPlatform/Java Card smart cards.

Computational Identification of Protein-Coding Sequences by Comparative Analysis

Gene prediction is an essential step in understanding the genome of a species once it has been sequenced. For that, a promising direction in current research on gene finding is a comparative genomics approach. In this paper, we present a novel approach to identifying evolutionarily conserved protein-coding sequences in genomes. The method takes advantage of the specific substitution pattern of coding se- quences together with the consistency of reading frames. It has been implemented in a software called Protea. Large- scale experimentation shows good results. Protea is in- tended to be a useful complement to existing tools based on homology search or statistical properties of the sequences.

Verifiable control flow policies for java bytecode

This paper presents the enforcement of control flow policies for Java bytecode dedicated to open and constrained devices. On-device enforcement of security policies mostly relies on run-time monitoring or inline checking code, which is not appropriate for strongly constrained devices such as mobile phones and smart-cards. We present a proof-carrying code approach with on-device lightweight verification of control flow policies statically at loading-time. Our approach is suitable for evolving, open and constrained Java-based systems as it is compositional, to avoid re-verification of already verified bytecode upon loading of new bytecode, and it is regressive, to cleanly support bytecode unloading.