Haplotype-resolved genome assembly and implementation of VitExpress, an open interactive transcriptomic platform for grapevine.

Haplotype-resolved genome assemblies were produced for Chasselas and Ugni Blanc, two heterozygous Vitis vinifera cultivars by combining high-fidelity long-read sequencing and high-throughput chromosome conformation capture (Hi-C). The telomere-to-telomere full coverage of the chromosomes allowed us to assemble separately the two haplo-genomes of both cultivars and revealed structural variations between the two haplotypes of a given cultivar. The deletions/insertions, inversions, translocations, and duplications provide insight into the evolutionary history and parental relationship among grape varieties. Integration of de novo single long-read sequencing of full-length transcript isoforms (Iso-Seq) yielded a highly improved genome annotation. Given its higher contiguity, and the robustness of the IsoSeq-based annotation, the Chasselas assembly meets the standard to become the annotated reference genome for V. vinifera. Building on these resources, we developed VitExpress, an open interactive transcriptomic platform, that provides a genome browser and integrated web tools for expression profiling, and a set of statistical tools (StatTools) for the identification of highly correlated genes. Implementation of the correlation finder tool for MybA1, a major regulator of the anthocyanin pathway, identified candidate genes associated with anthocyanin metabolism, whose expression patterns were experimentally validated as discriminating between black and white grapes. These resources and innovative tools for mining genome-related data are anticipated to foster advances in several areas of grapevine research.

Ploidy-specific transcriptomes shed light on the heterogeneous identity and metabolism of developing tomato pericarp cells.

Endoreduplication, during which cells increase their DNA content through successive rounds of full genome replication without cell division, is the major source of endopolyploidy in higher plants. Endoreduplication plays pivotal roles in plant growth and development and is associated with the activation of specific transcriptional programmes that are characteristic of each cell type, thereby defining their identity. In plants, endoreduplication is found in numerous organs and cell types, especially in agronomically valuable ones, such as the fleshy fruit (pericarp) of tomato presenting high ploidy levels. We used the tomato pericarp tissue as a model system to explore the transcriptomes associated with endoreduplication progression during fruit growth. We confirmed that expression globally scales with ploidy level and identified sets of differentially expressed genes presenting only developmental-specific, only ploidy-specific expression patterns or profiles resulting from an additive effect of ploidy and development. When comparing ploidy levels at a specific developmental stage, we found that non-endoreduplicated cells are defined by cell division state and cuticle synthesis while endoreduplicated cells are mainly defined by their metabolic activity changing rapidly over time. By combining this dataset with publicly available spatiotemporal pericarp expression data, we proposed a map describing the distribution of ploidy levels within the pericarp. These transcriptome-based predictions were validated by quantifying ploidy levels within the pericarp tissue. This in situ ploidy quantification revealed the dynamic progression of endoreduplication and its cell layer specificity during early fruit development. In summary, the study sheds light on the complex relationship between endoreduplication, cell differentiation and gene expression patterns in the tomato pericarp.

Transition to ripening in tomato requires hormone-controlled genetic reprogramming initiated in gel tissue.

Ripening is the last stage of the developmental program in fleshy fruits. During this phase, fruits become edible and acquire their unique sensory qualities and post-harvest potential. Although our knowledge of the mechanisms that regulate fruit ripening has improved considerably over the past decades, the processes that trigger the transition to ripening remain poorly deciphered. While transcriptomic profiling of tomato (Solanum lycopersicum L.) fruit ripening to date has mainly focused on the changes occurring in pericarp tissues between the Mature Green and Breaker stages, our study addresses the changes between the Early Mature Green and Late Mature Green stages in the gel and pericarp separately. The data showed that the shift from an inability to initiate ripening to the capacity to undergo full ripening requires extensive transcriptomic reprogramming that takes place first in the locular tissues before extending to the pericarp. Genome-wide transcriptomic profiling revealed the wide diversity of transcription factor (TF) families engaged in the global reprogramming of gene expression and identified those specifically regulated at the Mature Green stage in the gel but not in the pericarp, thereby providing potential targets toward deciphering the initial factors and events that trigger the transition to ripening. The study also uncovered an extensive reformed homeostasis for most plant hormones, highlighting the multihormonal control of ripening initiation. Our data unveil the antagonistic roles of ethylene and auxin during the onset of ripening and show that auxin treatment delays fruit ripening via impairing the expression of genes required for System-2 autocatalytic ethylene production that is essential for climacteric ripening. This study unveils the detailed features of the transcriptomic reprogramming associated with the transition to ripening of tomato fruit and shows that the first changes occur in the locular gel before extending to pericarp and that a reformed auxin homeostasis is essential for the ripening to proceed.

Ethylene signaling modulates tomato pollen tube growth through modifications of cell wall remodeling and calcium gradient.

Ethylene modulates plant developmental processes including flower development. Previous studies have suggested ethylene participates in pollen tube (PT) elongation, and both ethylene production and perception seem critical at the time of fertilization. The full gene set regulated by ethylene during PT growth is unknown. To study this, we used various EThylene Receptor (ETR) tomato (Solanum lycopersicum) mutants: etr3-ko, a loss-of-function (LOF) mutant; and NR (NEVER RIPE), a gain-of-function (GOF) mutant. The etr3-ko PTs grew faster than wild-type (WT) PTs. Oppositely, NR PT elongation was slower than in WT, and PTs displayed larger diameters. ETR mutations result in feedback control of ethylene production. Furthermore, ethylene treatment of germinating pollen grains increased PT length in etr-ko mutants and WT, but not in NR. Treatment with the ethylene perception inhibitor 1-methylcyclopropene decreased PT length in etr-ko mutants and WT, but had no effect on NR. This confirmed that ethylene regulates PT growth. The comparison of PT transcriptomes in LOF and GOF mutants, etr3-ko and NR, both harboring mutations of the ETR3 gene, revealed that ethylene perception has major impacts on cell wall- and calcium-related genes as confirmed by microscopic observations showing a modified distribution of the methylesterified homogalacturonan pectic motif and of calcium load. Our results establish links between PT growth, ethylene, calcium, and cell wall metabolism, and also constitute a transcriptomic resource.

Histone posttranslational modifications rather than DNA methylation underlie gene reprogramming in pollination-dependent and pollination-independent fruit set in tomato.

Fruit formation comprises a series of developmental transitions among which the fruit set process is essential in determining crop yield. Yet, our understanding of the epigenetic landscape remodelling associated with the flower-to-fruit transition remains poor. We investigated the epigenetic and transcriptomic reprogramming underlying pollination-dependent and auxin-induced flower-to-fruit transitions in the tomato (Solanum lycopersicum) using combined genomewide transcriptomic profiling, global ChIP-sequencing and whole genomic DNA bisulfite sequencing (WGBS). Variation in the expression of the overwhelming majority of genes was associated with change in histone mark distribution, whereas changes in DNA methylation concerned a minor fraction of differentially expressed genes. Reprogramming of genes involved in processes instrumental to fruit set correlated with their H3K9ac or H3K4me3 marking status but not with changes in cytosine methylation, indicating that histone posttranslational modifications rather than DNA methylation are associated with the remodelling of the epigenetic landscape underpinning the flower-to-fruit transition. Given the prominent role previously assigned to DNA methylation in reprogramming key genes of the transition to ripening, the outcome of the present study supports the idea that the two main developmental transitions in fleshy fruit and the underlying transcriptomic reprogramming are associated with different modes of epigenetic regulations.

Transcriptome profiling of sorted endoreduplicated nuclei from tomato fruits: how the global shift in expression ascribed to DNA ploidy influences RNA-Seq data normalization and interpretation.

As part of normal development most eukaryotic organisms, ranging from insects and mammals to plants, display variations in nuclear ploidy levels resulting from somatic endopolyploidy. Endoreduplication is the major source of endopolyploidy in higher plants. Endoreduplication is a remarkable characteristic of the fleshy pericarp tissue of developing tomato fruits, where it establishes a highly integrated cellular system that acts as a morphogenetic factor supporting cell growth. However, the functional significance of endoreduplication is not fully understood. Although endoreduplication is thought to increase metabolic activity due to a global increase in transcription, the issue of gene-specific ploidy-regulated transcription remains open. To investigate the influence of endoreduplication on transcription in tomato fruit, we tested the feasibility of a RNA sequencing (RNA-Seq) approach using total nuclear RNA extracted from purified populations of flow cytometry-sorted nuclei based on their DNA content. Here we show that cell-based approaches to the study of RNA-Seq profiles need to take into account the putative global shift in expression between samples for correct analysis and interpretation of the data. From ploidy-specific expression profiles we found that the activity of cells inside the pericarp is related both to the ploidy level and their tissue location.

TomExpress, a unified tomato RNA-Seq platform for visualization of expression data, clustering and correlation networks.

The TomExpress platform was developed to provide the tomato research community with a browser and integrated web tools for public RNA-Seq data visualization and data mining. To avoid major biases that can result from the use of different mapping and statistical processing methods, RNA-Seq raw sequence data available in public databases were mapped de novo on a unique tomato reference genome sequence and post-processed using the same pipeline with accurate parameters. Following the calculation of the number of counts per gene in each RNA-Seq sample, a communal global normalization method was applied to all expression values. This unifies the whole set of expression data and makes them comparable. A database was designed where each expression value is associated with corresponding experimental annotations. Sample details were manually curated to be easily understandable by biologists. To make the data easily searchable, a user-friendly web interface was developed that provides versatile data mining web tools via on-the-fly generation of output graphics, such as expression bar plots, comprehensive in planta representations and heatmaps of hierarchically clustered expression data. In addition, it allows for the identification of co-expressed genes and the visualization of correlation networks of co-regulated gene groups. TomExpress provides one of the most complete free resources of publicly available tomato RNA-Seq data, and allows for the immediate interrogation of transcriptional programs that regulate vegetative and reproductive development in tomato under diverse conditions. The design of the pipeline developed in this project enables easy updating of the database with newly published RNA-Seq data, thereby allowing for continuous enrichment of the resource.

Comprehensive Profiling of Ethylene Response Factor Expression Identifies Ripening-Associated ERF Genes and Their Link to Key Regulators of Fruit Ripening in Tomato.

Our knowledge of the factors mediating ethylene-dependent ripening of climacteric fruit remains limited. The transcription of ethylene-regulated genes is mediated by ethylene response factors (ERFs), but mutants providing information on the specific role of the ERFs in fruit ripening are still lacking, likely due to functional redundancy among this large multigene family of transcription factors. We present here a comprehensive expression profiling of tomato (Solanum lycopersicum) ERFs in wild-type and tomato ripening-impaired tomato mutants (Never-ripe [Nr], ripening-inhibitor [rin], and non-ripening [nor]), indicating that out of the 77 ERFs present in the tomato genome, 27 show enhanced expression at the onset of ripening while 28 display a ripening-associated decrease in expression, suggesting that different ERFs may have contrasting roles in fruit ripening. Among the 19 ERFs exhibiting the most consistent up-regulation during ripening, the expression of 11 ERFs is strongly down-regulated in rin, nor, and Nr tomato ripening mutants, while only three are consistently up-regulated. Members of subclass E, SlERF.E1, SlERF.E2, and SlERF.E4, show dramatic down-regulation in the ripening mutants, suggesting that their expression might be instrumental in fruit ripening. This study illustrates the high complexity of the regulatory network connecting RIN and ERFs and identifies subclass E members as the most active ERFs in ethylene- and RIN/NOR-dependent ripening.

Overexpression of the class D MADS-box gene Sl-AGL11 impacts fleshy tissue differentiation and structure in tomato fruits.

MADS-box transcription factors are key elements of the genetic networks controlling flower and fruit development. Among these, the class D clade gathers AGAMOUS-like genes which are involved in seed, ovule, and funiculus development. The tomato genome comprises two class D genes, Sl-AGL11 and Sl-MBP3, both displaying high expression levels in seeds and in central tissues of young fruits. The potential effects of Sl-AGL11 on fruit development were addressed through RNAi silencing and ectopic expression strategies. Sl-AGL11-down-regulated tomato lines failed to show obvious phenotypes except a slight reduction in seed size. In contrast, Sl-AGL11 overexpression triggered dramatic modifications of flower and fruit structure that include: the conversion of sepals into fleshy organs undergoing ethylene-dependent ripening, a placenta hypertrophy to the detriment of locular space, starch and sugar accumulation, and an extreme softening that occurs well before the onset of ripening. RNA-Seq transcriptomic profiling highlighted substantial metabolic reprogramming occurring in sepals and fruits, with major impacts on cell wall-related genes. While several Sl-AGL11-related phenotypes are reminiscent of class C MADS-box genes (TAG1 and TAGL1), the modifications observed on the placenta and cell wall and the Sl-AGL11 expression pattern suggest an action of this class D MADS-box factor on early fleshy fruit development.

Comprehensive transcript profiling of two grapevine rootstock genotypes contrasting in drought susceptibility links the phenylpropanoid pathway to enhanced tolerance.

In light of ongoing climate changes in wine-growing regions, the selection of drought-tolerant rootstocks is becoming a crucial factor for developing a sustainable viticulture. In this study, M4, a new rootstock genotype that shows tolerance to drought, was compared from a genomic and transcriptomic point of view with the less drought-tolerant genotype 101.14. The root and leaf transcriptome of both 101.14 and the M4 rootstock genotype was analysed, following exposure to progressive drought conditions. Multifactorial analyses indicated that stress treatment represents the main factor driving differential gene expression in roots, whereas in leaves the genotype is the prominent factor. Upon stress, M4 roots and leaves showed a higher induction of resveratrol and flavonoid biosynthetic genes, respectively. The higher expression of VvSTS genes in M4, confirmed by the accumulation of higher levels of resveratrol in M4 roots compared with 101.14, was coupled to an up-regulation of several VvWRKY transcription factors. Interestingly, VvSTS promoter analyses performed on both the resequenced genomes highlighted a significantly higher number of W-BOX elements in the tolerant genotype. It is proposed that the elevated synthesis of resveratrol in M4 roots upon water stress could enhance the plant's ability to cope with the oxidative stress usually associated with water deficit.

Genome-wide identification, phylogenetic analysis, expression profiling, and protein-protein interaction properties of TOPLESS gene family members in tomato.

Members of the TOPLESS gene family emerged recently as key players in gene repression in several mechanisms, especially in auxin perception. The TOPLESS genes constitute, in 'higher-plant' genomes, a small multigenic family comprising four to 11 members. In this study, this family was investigated in tomato, a model plant for Solanaceae species and fleshy fruits. Six open reading frames predicted to encode topless-like proteins (SlTPLs) containing the canonical domains (LisH, CTLH, and two WD40 repeats) were identified in the tomato genome. Nuclear localization was confirmed for all members of the SlTPL family with the exception SlTPL6, which localized at the cytoplasm and was excluded from the nucleus. SlTPL genes displayed distinctive expression patterns in different tomato organs, with SlTPL1 showing the highest levels of transcript accumulation in all tissues tested except in ripening fruit where SlTPL3 and SlTPL4 were the most prominently expressed. To gain insight into the specificity of the different TOPLESS paralogues, a protein-protein interaction map between TOPLESS and auxin/indole-3-acetic acid (Aux/IAA) proteins was built using a yeast two-hybrid approach. The PPI map enabled the distinction of two patterns: TOPLESS isoforms interacting with the majority of Aux/IAA, and isoforms with limited capacity for interaction with these protein partners. Interestingly, evolutionary analyses of the TOPLESS gene family revealed that the highly expressed isoforms (SlTPL1, SlTPL3, and SlTPL4) corresponded to the three TPL-related genes undergoing the strongest purifying selection, while the selection was much weaker for SlTPL6, which was expressed at a low level and encoded a protein lacking the capacity to interact with Aux/IAAs.

Proteomic analysis of chloroplast-to-chromoplast transition in tomato reveals metabolic shifts coupled with disrupted thylakoid biogenesis machinery and elevated energy-production components.

A comparative proteomic approach was performed to identify differentially expressed proteins in plastids at three stages of tomato (Solanum lycopersicum) fruit ripening (mature-green, breaker, red). Stringent curation and processing of the data from three independent replicates identified 1,932 proteins among which 1,529 were quantified by spectral counting. The quantification procedures have been subsequently validated by immunoblot analysis of six proteins representative of distinct metabolic or regulatory pathways. Among the main features of the chloroplast-to-chromoplast transition revealed by the study, chromoplastogenesis appears to be associated with major metabolic shifts: (1) strong decrease in abundance of proteins of light reactions (photosynthesis, Calvin cycle, photorespiration) and carbohydrate metabolism (starch synthesis/degradation), mostly between breaker and red stages and (2) increase in terpenoid biosynthesis (including carotenoids) and stress-response proteins (ascorbate-glutathione cycle, abiotic stress, redox, heat shock). These metabolic shifts are preceded by the accumulation of plastid-encoded acetyl Coenzyme A carboxylase D proteins accounting for the generation of a storage matrix that will accumulate carotenoids. Of particular note is the high abundance of proteins involved in providing energy and in metabolites import. Structural differentiation of the chromoplast is characterized by a sharp and continuous decrease of thylakoid proteins whereas envelope and stroma proteins remain remarkably stable. This is coincident with the disruption of the machinery for thylakoids and photosystem biogenesis (vesicular trafficking, provision of material for thylakoid biosynthesis, photosystems assembly) and the loss of the plastid division machinery. Altogether, the data provide new insights on the chromoplast differentiation process while enriching our knowledge of the plant plastid proteome.

Selenium Biofortification Impacts the Tomato Fruit Metabolome and Transcriptional Profile at Ripening.

In the present work, the effects of enriching tomatoes with selenium were studied in terms of physiological, metabolic, and molecular processes in the last stages of fruit development, particularly during ripening. A selenium concentration of 10 mg L-1 with sodium selenate and selenium nanoparticles was used in the spray treatments on the whole plants. No significant effects of selenium enrichment were detected in terms of ethylene production or color changes in the ripening fruit. However, selenium enrichment had an influence on both the primary and secondary metabolic processes and thus the biochemical composition of ripe tomatoes. Selenium decreased the amount of ß-carotene, increased the accumulation of naringenin and chlorogenic acid, and decreased the coumaric acid level. Selenium also affected the volatile organic compound profile, with changes in the level of specific apocarotenoid compounds, such as ß-ionone. These metabolomic changes may, to some extent, be due to the impact of selenium treatment on the transcription of genes involved in the metabolism of these compounds. RNA-seq analysis showed that the selenium application mostly impacted the expression of the genes involved in hormonal signaling, secondary metabolism, flavonoid biosynthesis, and glycosaminoglycan degradation.

The auxin-responsive transcription factor SlDOF9 regulates inflorescence and flower development in tomato.

Understanding the mechanisms underlying differentiation of inflorescence and flower meristems is essential towards enlarging our knowledge of reproductive organ formation and to open new prospects for improving yield traits. Here, we show that SlDOF9 is a new modulator of floral differentiation in tomato. CRISPR/Cas9 knockout strategy uncovered the role of SlDOF9 in controlling inflorescence meristem and floral meristem differentiation via the regulation of cell division genes and inflorescence architecture regulator LIN. Tomato dof9-KO lines have more flowers in both determinate and indeterminate cultivars and produce more fruit upon vibration-assisted fertilization. SlDOF9 regulates inflorescence development through an auxin-dependent ARF5-DOF9 module that seems to operate, at least in part, differently in Arabidopsis and tomato. Our findings add a new actor to the complex mechanisms underlying reproductive organ differentiation in flowering plants and provide leads towards addressing the diversity of factors controlling the transition to reproductive organs.

Interaction of two MADS-box genes leads to growth phenotype divergence of all-flesh type of tomatoes.

All-flesh tomato cultivars are devoid of locular gel and exhibit enhanced firmness and improved postharvest storage. Here, we show that SlMBP3 is a master regulator of locular tissue in tomato fruit and that a deletion at the gene locus underpins the All-flesh trait. Intriguingly, All-flesh varieties lack the deleterious phenotypes reported previously for SlMBP3 under-expressing lines and which preclude any potential commercial use. We resolve the causal factor for this phenotypic divergence through the discovery of a natural mutation at the SlAGL11 locus, a close homolog of SlMBP3. Misexpressing SlMBP3 impairs locular gel formation through massive transcriptomic reprogramming at initial phases of fruit development. SlMBP3 influences locule gel formation by controlling cell cycle and cell expansion genes, indicating that important components of fruit softening are determined at early pre-ripening stages. Our findings define potential breeding targets for improved texture in tomato and possibly other fleshy fruits.

Optimization of an RNA-Seq Differential Gene Expression Analysis Depending on Biological Replicate Number and Library Size.

RNA-Seq is a widely used technology that allows an efficient genome-wide quantification of gene expressions for, for example, differential expression (DE) analysis. After a brief review of the main issues, methods and tools related to the DE analysis of RNA-Seq data, this article focuses on the impact of both the replicate number and library size in such analyses. While the main drawback of previous relevant studies is the lack of generality, we conducted both an analysis of a two-condition experiment (with eight biological replicates per condition) to compare the results with previous benchmark studies, and a meta-analysis of 17 experiments with up to 18 biological conditions, eight biological replicates and 100 million (M) reads per sample. As a global trend, we concluded that the replicate number has a larger impact than the library size on the power of the DE analysis, except for low-expressed genes, for which both parameters seem to have the same impact. Our study also provides new insights for practitioners aiming to enhance their experimental designs. For instance, by analyzing both the sensitivity and specificity of the DE analysis, we showed that the optimal threshold to control the false discovery rate (FDR) is approximately 2-r, where r is the replicate number. Furthermore, we showed that the false positive rate (FPR) is rather well controlled by all three studied R packages: DESeq, DESeq2, and edgeR. We also analyzed the impact of both the replicate number and library size on gene ontology (GO) enrichment analysis. Interestingly, we concluded that increases in the replicate number and library size tend to enhance the sensitivity and specificity, respectively, of the GO analysis. Finally, we recommend to RNA-Seq practitioners the production of a pilot data set to strictly analyze the power of their experimental design, or the use of a public data set, which should be similar to the data set they will obtain. For individuals working on tomato research, on the basis of the meta-analysis, we recommend at least four biological replicates per condition and 20 M reads per sample to be almost sure of obtaining about 1000 DE genes if they exist.

In Papyro Comparison of TMM (edgeR), RLE (DESeq2), and MRN Normalization Methods for a Simple Two-Conditions-Without-Replicates RNA-Seq Experimental Design.

In the past 5 years, RNA-Seq has become a powerful tool in transcriptome analysis even though computational methods dedicated to the analysis of high-throughput sequencing data are yet to be standardized. It is, however, now commonly accepted that the choice of a normalization procedure is an important step in such a process, for example in differential gene expression analysis. The present article highlights the similarities between three normalization methods: TMM from edgeR R package, RLE from DESeq2 R package, and MRN. Both TMM and DESeq2 are widely used for differential gene expression analysis. This paper introduces properties that show when these three methods will give exactly the same results. These properties are proven mathematically and illustrated by performing in silico calculations on a given RNA-Seq data set.

Grapevine Rootstocks Differentially Affect the Rate of Ripening and Modulate Auxin-Related Genes in Cabernet Sauvignon Berries.

In modern viticulture, grafting commercial grapevine varieties on interspecific rootstocks is a common practice required for conferring resistance to many biotic and abiotic stresses. Nevertheless, the use of rootstocks to gain these essential traits is also known to impact grape berry development and quality, although the underlying mechanisms are still poorly understood. In grape berries, the onset of ripening (véraison) is regulated by a complex network of mobile signals including hormones such as auxins, ethylene, abscisic acid, and brassinosteroids. Recently, a new rootstock, designated M4, was selected based on its enhanced tolerance to water stress and medium vigor. This study investigates the effect of M4 on Cabernet Sauvignon (CS) berry development in comparison to the commercial 1103P rootstock. Physical and biochemical parameters showed that the ripening rate of CS berries is faster when grafted onto M4. A multifactorial analysis performed on mRNA-Seq data obtained from skin and pulp of berries grown in both graft combinations revealed that genes controlling auxin action (ARF and Aux/IAA) represent one of main categories affected by the rootstock genotype. Considering that the level of auxin tightly regulates the transcription of these genes, we investigated the behavior of the main gene families involved in auxin biosynthesis and conjugation. Molecular and biochemical analyses confirmed a link between the rate of berry development and the modulation of auxin metabolism. Moreover, the data indicate that this phenomenon appears to be particularly pronounced in skin tissue in comparison to the flesh.

Statistical properties of the quantile normalization method for density curve alignment.

The article investigates the large sample properties of the quantile normalization method by Bolstad et al. (2003) [4] which has become one of the most popular methods to align density curves in microarray data analysis. We prove consistency of this method which is viewed as a particular case of the structural expectation procedure for curve alignment, which corresponds to a notion of barycenter of measures in the Wasserstein space. Moreover, we show that, this method fails in some case of mixtures, and we propose a new methodology to cope with this issue.

Comparison of normalization methods for differential gene expression analysis in RNA-Seq experiments: A matter of relative size of studied transcriptomes.

In recent years, RNA-Seq technologies became a powerful tool for transcriptome studies. However, computational methods dedicated to the analysis of high-throughput sequencing data are yet to be standardized. In particular, it is known that the choice of a normalization procedure leads to a great variability in results of differential gene expression analysis. The present study compares the most widespread normalization procedures and proposes a novel one aiming at removing an inherent bias of studied transcriptomes related to their relative size. Comparisons of the normalization procedures are performed on real and simulated data sets. Real RNA-Seq data sets analyses, performed with all the different normalization methods, show that only 50% of significantly differentially expressed genes are common. This result highlights the influence of the normalization step on the differential expression analysis. Real and simulated data sets analyses give similar results showing 3 different groups of procedures having the same behavior. The group including the novel method named "Median Ratio Normalization" (MRN) gives the lower number of false discoveries. Within this group the MRN method is less sensitive to the modification of parameters related to the relative size of transcriptomes such as the number of down- and upregulated genes and the gene expression levels. The newly proposed MRN method efficiently deals with intrinsic bias resulting from relative size of studied transcriptomes. Validation with real and simulated data sets confirmed that MRN is more consistent and robust than existing methods.