Genetic determinants of antibiotic resistance (AR) have been extensively investigated. High-throughput sequencing allows for the assessment of the relationship between genotype and phenotype. A panel of 672 Pseudomonas aeruginosa strains was analysed, including representatives of globally disseminated multidrug-resistant and extensively drug-resistant clones; genomes and multiple antibiograms were available. This panel was annotated for AR gene presence and polymorphism, defining a resistome in which integrons were included. Integrons were present in >70 distinct cassettes, with In5 being the most prevalent. Some cassettes closely associated with clonal complexes, whereas others spread across the phylogenetic diversity, highlighting the importance of horizontal transfer. A resistome-wide association study (RWAS) was performed for clinically relevant antibiotics by correlating the variability in minimum inhibitory concentration (MIC) values with resistome data. Resistome annotation identified 147 loci associated with AR. These loci consisted mainly of acquired genomic elements and intrinsic genes. The RWAS allowed for correct identification of resistance mechanisms for meropenem, amikacin, levofloxacin and cefepime, and added 46 novel mutations. Among these, 29 were variants of the oprD gene associated with variation in meropenem MIC. Using genomic and MIC data, phenotypic AR was successfully correlated with molecular determinants at the whole-genome sequence level.

The majority of gene expression studies focus on the search for genes whose mean expression is different between two or more populations of samples in the so-called “differential expression analysis” approach. However, a difference in variance in gene expression may also be biologically and physiologically relevant. In the classical statistical model used to analyze RNA-sequencing (RNA-seq) data, the dispersion, which defines the variance, is only considered as a parameter to be estimated prior to identifying a difference in mean expression between conditions of interest. Here, we propose to evaluate four recently published methods, which detect differences in both the mean and dispersion in RNA-seq data. We thoroughly investigated the performance of these methods on simulated datasets and characterized parameter settings to reliably detect genes with a differential expression dispersion. We applied these methods to The Cancer Genome Atlas datasets. Interestingly, among the genes with an increased expression dispersion in tumors and without a change in mean expression, we identified some key cellular functions, most of which were related to catabolism and were overrepresented in most of the analyzed cancers. In particular, our results highlight autophagy, whose role in cancerogenesis is context-dependent, illustrating the potential of the differential dispersion approach to gain new insights into biological processes and to discover new biomarkers.

Background:One of the first steps of a usual RNA-seq data analysis workflow consists in quantifying gene expression by aligning the sequencing reads to a reference genome and counting the aligned reads in its annotated regions. Downstream analysis, such as the identification of differentially expressed genes, strongly rely on the quality of this process. In addition to the performance of alignment methods, the choice of reference genome, when several of them are available, may strongly impact this step.Results:Here, we propose to evaluate the effect of widely used \Rn reference genomes consisting of Ensembl and RefSeq annotations of the Rnor\_6.0 assembly and the ones based on the mRatBN7.2 genome assembly recently published by RefSeq, on a classical differential expression workflow. We re-analyzed published RNA-seq datasets from different hippocampal subregions and revealed that the Ensembl and Refseq reference genome based on the mRatBN7.2 assembly provide an improvement of read mapping statistics. We showed that the RefSeq annotations of this assembly make gene expression quantification and differentially expressed gene identification more reliable thanks to overall longer exon length in comparison with Ensembl annotations. Moreover, we identified specific biologically relevant results using RefSeq annotations of the new genome assembly.Conclusion:Overall, the biological interpretation of the differential expression analysis of the analyzed datasets may be dramatically impacted by the choice of reference genome. Therefore, we believe that this choice should be more carefully addressed and that our approach could extend to other tissues and species.

The majority of gene expression studies focus on the search for genes whose mean expression is different between two or more populations of samples in the so-called "differential expression analysis" approach. However, a difference in variance in gene expression may also be biologically and physiologically relevant. In the classical statistical model used to analyze RNA-sequencing (RNA-seq) data, the dispersion, which defines the variance, is only considered as a parameter to be estimated prior to identifying a difference in mean expression between conditions of interest. Here, we propose to evaluate two recent methods, MDSeq and DiPhiSeq, which detect differences in both the mean and dispersion in RNA-seq data. We thoroughly investigated the performance of these methods on simulated datasets and characterized parameter settings to reliably detect genes with a differential expression dispersion. We applied both methods to The Cancer Genome Atlas datasets. Interestingly, among the genes with an increased expression dispersion in tumors and without a change in mean expression, we identified some key cellular functions, most of which were related to catabolism and were overrepresented in most of the analyzed cancers. In particular, our results highlight autophagy, whose role in cancerogenesis is context-dependent, illustrating the potential of the differential dispersion approach to gain new insights into biological processes.