13h30: Marie-Laure Martin Magniette (URGV/UMR518, INRA/AgroParisTech)
HTSDiff : a Poisson mixture model for differential gene expression analysis of RNA-seq data
Based on their ability to detect and count individual cDNA molecules, next generation sequencing technologies are promising to radically change the way transcriptomic analyses are performed. To provide a critical assessment of this claim in plant gene expression analysis, the URGV performed a direct comparison of RNA-seq with the latest high-density CATMA microarray using Arabidopsis thaliana samples and validated more than 300 genes by qPCR on the same samples.
Surprisingly, CATMA microarrays largely outperformed RNA-seq in terms of the number of differentially expressed genes detected in several comparisons. A detailed analysis strongly suggested that, although RNA-seq data are potentially better at quantifying gene expression than microarray data, they are not optimally exploited by the statistical methods currently used for differential analysis. To tackle this problem, we recast the comparison of two samples as an unsupervised classification problem based on a mixture model of Poisson distributions. Methods are compared and discussed according to the qPCR data and synthetic datasets.
14h30: Lydia Robert (INRA/UPMC)
Division control in bacteria
Many organisms coordinate cell growth and division through size control mechanisms: cells must reach a critical size to trigger some cell cycle event. Bacterial division is often assumed to be under such control, but definite evidence is still lacking. Deciding whether division control relies on a "timer" or "sizer" mechanism requires quantitative comparisons between models and data. The "timer" and "sizer" hypotheses find a natural translation in models based on Partial Differential Equations. We confronted these models with recent data on Escherichia coli single cell growth. We demonstrated that a size-independent "timer" mechanism for division control, though theoretically possible, is quantitatively incompatible with the data and extremely sensitive to slight variations in the growth law. In contrast, a "sizer" model is robust and fits the data well.