ABSTRACT. Abstract

Invasive fungal diseases such as candidiasis, caused by Candida species, are a major public health problem. They are difficult to diagnose and have high mortality rates. Moreover therapeutic options are limited, and resistance to multiple antifungal drugs is increasingly reported, particularly in emerging species such as Candida glabrata, Candida parapsilosis, and Candida auris. Despite their common genus name, Candida pathogens, are evolutionarily diverse and belong to different lineages where the ability to infect humans has emerged independently. Over the last years we have used in vitro evolution, and comparative genomics approaches to understand how the different pathogenic lineages adapted to humans and how they become resistant to the drugs we use to fight their infections.

ABSTRACT. The chronology and phylogeny of bacterial evolution are difficult to reconstruct due to a scarce fossil record. The analysis of bacterial genomes remains challenging because of large sequence divergence, the plasticity of bacterial genomes due to frequent gene loss, horizontal gene transfer, and differences in selective pressure from one locus to another. Therefore, taking advantage of the rich and rapidly accumulating genomic data requires accurate modeling of genome evolution. An important technical consideration is that loci with high effective mutation rates may diverge beyond the detection limit of the alignment algorithms used, biasing the genome-wide divergence estimates toward smaller divergences. In this article, we propose a novel method to gain insight into bacterial evolution based on statistical properties of genome comparisons. We find that the length distribution of sequence matches is shaped by the effective mutation rates of different loci, by the horizontal transfers, and by the aligner sensitivity. Based on these inputs, we build a model and show that it accounts for the empirically observed distributions, taking the Enterobacteriaceae family as an example. Our method allows to distinguish segments of vertical and horizontal origins and to estimate the time divergence and exchange rate between any pair of taxa from genome-wide alignments. Based on the estimated time divergences, we construct a time-calibrated phylogenetic tree to demonstrate the accuracy of the method.

ABSTRACT. Molecular evolution is known to proceed in a non-homogenous manner in protein sequences. For instance, amino-acid substitutions occur at different rates depending on the associated gene and on their location inside the sequence. Additionally all amino acids changes are not equally likely to happen. We hypothetise that part of the evolution of a protein sequence can be predicted, either due to general statistical bias or more specific and yet unknown rules. In this study we test this hypothesis by training a deep neural network model, DeepEvoSeq, built on AlphaFold, to predict the evolution of protein sequences. We design a specific experiment where we aim to predict amino-acid sequences of a target species given orthologous sequences of closely related species. To this end we split the prediction of the evolution in two distinct tasks: the prediction of the positions of substitutions along the protein sequences, and the prediction of the nature of the amino acid change given the positions of substitutions. We show that a large part of the evolution in protein sequences is predictable, and that the nature of amino acid changes is more predictable than the positions of substitutions.

ABSTRACT. Several studies showed that folds (topology of protein secondary structures) distribution in proteomes may be a global proxy to build phylogeny. Then, some folds should be synapomorphies (derived characters exclusively shared among taxa). However, previous studies used methods that did not allow synapomorphy identification, which requires congruence analysis of folds as individual characters. Here, we map SCOP folds onto a sample of 210 species across the tree of life (TOL). Congruence is assessed using retention index of each fold for the TOL, and principal component analysis for deeper branches. Using a bicluster mapping approach we define synapomorphic blocks of folds (SBF) sharing similar presence/absence patterns. Among the 1,232 folds, 20% are universally present in our TOL, while 54% are reliable synapomorphies. These results are similar with CATH and ECOD databases. Eukaryotes are characterized by a large number of them, and several SBFs clearly supported nested eukaryotic clades (divergence times from 1,100 to 380 mya). While clearly separated, the three superkingdoms reveal a strong mosaic pattern. This pattern is consistent with the dual origin of eukaryotes and witness secondary endosymbiosis in their phothosynthetic clades. Eukaryotes have more protein folds than archaea and bacteria. These folds are of two types: shared with archaea and/or bacteria on one hand and specific to eukaryotic clades on the other hand. The first kind of folds is inherited from the first endosymbiosis and confirms the mixed origin of eukaryotes. We have identified 28 eukaryotic folds unambiguously inherited from bacteria and 40 eukaryotic folds unambiguously inherited from archaea. Compared to previous studies, the repartition of informational function is higher than expected for folds originated from Bacteria and as high as expected for folds inherited from Archaea. The second type of folds is specifically eukaryotic and associated with an increase of new folds within eukaryotes, distributed in particular clades. Reconstructed ancestral states coupled with dating of each node on the tree of life provided fold appearance rates. The rate is on average twice higher within Eukaryota than within Bacteria or Archaea. The highest rates are found in the origins of eukaryotes, holozoans, metazoans, metazoans stricto sensu and vertebrates: the roots of these clades correspond to bursts of fold evolution. We could correlate the functions of some of the fold synapomorphies within eukaryotes with significant evolutionary events. Among them, we find evidence for the rise of multicellularity, adaptive immune system or virus folds which could be linked to an ecological shift made by tetrapods.

ABSTRACT. Long-read RNA sequencing has revolutionized genome annotation. Here, we present insights from the IGDRion sequencing Facility (https://igdr.univ-rennes.fr/en/igdrion) located in Rennes. IGDRion is equipped with cutting-edge long-read sequencing devices and bioinformatic tools. We will present our nextflow-based pipeline, ANNEXA, which facilitates comprehensive analysis of coding and non-coding transcripts (lncRNAs). We applied ANNEXA to human and canine cancer cell lines and identified hundreds of new genes and alternative isoforms. In conclusion, the IGDRion platform provides a robust infrastructure for advancing transcriptomic research with long-read sequencing technologies.

ABSTRACT. Background Initially developed for the analysis of the genome of Acinetobacter baylyi ADP1 in 2002, the Microscope platform was launched in 2005. Since then, it has been under continuous development to meet the needs of microbiologists in a context of rapidly evolving sequencing technologies. Over the years, the platform has been widely used by microbiologists from academia and industry all around the world for collaborative studies and expert annotation.

Results MicroScope is an integrated Web platform for management, annotation, comparative analysis and visualization of microbial genomes (https://mage.genoscope.cns.fr/microscope) [1]. This platform facilitates collaborative research within a comprehensive comparative genomics framework, thereby enhancing community-driven curation initiatives. MicroScope provides analyses for complete and ongoing genome projects together with metabolic network reconstruction and transcriptomic experiments allowing users to improve the understanding of gene functions. The platform integrates more than 30 workflows to perform structural and functional annotation. They integrate tools and databases developed in academia that allow analyzing a wide range of biological systems (antibiotic resistance, secondary metabolites, secretions systems, defense systems,...). The platform also has extensive functionalities to explore and compare metabolic pathways. Recent functionalities allow users to perform comparative pangenomics on hundreds of genomes of the same species and to explore their content in regions of genomic plasticity.

Conclusions MicroScope has been described in 8 publication and book chapters over the years. To date, the platform contains data for >20,000 microbial genomes, part of which are manually curated and maintained by an international community of microbiologists (>7,100 user accounts in March 2024 among which only 35% from France). We offer professional training, but the platform is also a useful resource for academic training.

References 1. Vallenet, David, Alexandra Calteau, Mathieu Dubois, Paul Amours, Adelme Bazin, Mylène Beuvin, Laura Burlot, et al. 2020. “MicroScope: An Integrated Platform for the Annotation and Exploration of Microbial Gene Functions through Genomic, Pangenomic and Metabolic Comparative Analysis.” Nucleic Acids Research 48 (D1): D579–89.

ABSTRACT. Background: Alternative splicing is a pivotal mechanism contributing to transcriptome complexity and proteome diversity, influencing key biological processes and linked to various diseases. Single-cell RNA sequencing (scRNA-seq) using long-read technologies, such as Oxford Nanopore Technologies (ONT), offers a promising approach to resolve the isoform complexity at a single-cell level, overcoming limitations of short-read sequencing methods which lose valuable isoform information.

Methods: We benchmarked seven bioinformatics pipelines for processing scRNA-seq data generated by Nanopore sequencing technology, including hybrid methods (Sicelore1, Snuupy2, scNapBar3) and Nanopore-only methods (FLAMES1, Sockeye5, Sicelore2.11, scNanoGPS6). Our evaluation encompasses their performance across multiple datasets, including scRNA-seq Illumina, ONT MinION, ONT PromethION, ONT Visium spatial, and simulated data, assessing metrics such as speed and memory usage, scalability, precision and recall on the barcode and UMI correction, transcripts quantification and biological signal integrity at the gene and isoform levels.

Results: Hybrid approaches, while offering high precision in cell barcode assignment, were hampered by sequencing cost and complexity of bioinformatics analysis. Nanopore-only methods, particularly Sicelore 2.1 and Sockeye, presented an efficient compromise in terms of read assignment accuracy and computational resources. These methods demonstrated the ability to detect and quantify isoforms correctly, with a significant correlation in gene expression levels compared to Illumina data. Notably, FLAMES tended to overestimate UMI and gene counts, impacting negatively the accuracy of gene expression estimates. Analysis also revealed the impact of sequencing depth on the correlation with Illumina data, highlighting the trade-offs in choosing between different sequencing and analysis strategies.

Conclusion: Our benchmark of bioinformatics tools for scRNA-seq Nanopore long-read data underscores the importance of selecting appropriate methods based on specific study goals and data characteristics. While hybrid methods provide high precision, Nanopore-only approaches offer a balanced solution for effective isoform resolution in single-cell studies.

Acknowledgements The IBENS genomics core facility was supported by the France Génomique national infrastructure, funded as part of the “Investissements d'Avenir” program managed by the Agence Nationale de la Recherche (contract ANR-10-INBS-09). This work was supported by Institut Universitaire de France and ITMO Cancer grant N° 21CD032-00.

References: 1. Lebrigand K, Magnone V, Barbry P, Waldmann R. High throughput error corrected Nanopore single cell transcriptome sequencing. Nat Commun. 12 août 2020;11(1):4025. 2. FlsnRNA-seq: protoplasting-free full-length single-nucleus RNA profiling in plants | Genome Biology | Full Text [Internet]. [cité 5 mai 2023]. Disponible sur: https://genomebiology.biomedcentral.com/articles/10.1186/s13059-021-02288-0 3. Wang Q, Bönigk S, Böhm V, Gehring N, Altmüller J, Dieterich C. Single-cell transcriptome sequencing on the Nanopore platform with ScNapBar. RNA. juill 2021;27(7):763‑70. 4. Tian L, Jabbari JS, Thijssen R, Gouil Q, Amarasinghe SL, Voogd O, et al. Comprehensive characterization of single-cell full-length isoforms in human and mouse with long-read sequencing. Genome Biol. 11 nov 2021;22(1):310. 5. nanoporetech/sockeye [Internet]. Oxford Nanopore Technologies; 2023 [cité 8 févr 2024]. Disponible sur: https://github.com/nanoporetech/sockeye 6. Shiau CK, Lu L, Kieser R, Fukumura K, Pan T, Lin HY, et al. High throughput single cell long-read sequencing analyses of same-cell genotypes and phenotypes in human tumors. Nat Commun. 11 juill 2023;14(1):4124.

ABSTRACT. DNA methylation is an epigenetic mark involved in the regulation of gene expression and patterns of DNA methylation anticorrelates with chromatin accessibility and transcription factor binding. DNA methylation can be profiled at the single cytosine resolution in the whole genome and has been performed in many cell types and conditions. Computational approaches are then essential to study DNA methylation patterns in a single condition or capture dynamic changes of DNA methylation levels across conditions. Towards this goal, we developed MethyLasso, a new approach based on the segmentation of DNA methylation data, that enables the identification of low-methylated regions (LMRs), unmethylated regions (UMRs), DNA methylation valleys (DMVs) and partially methylated domains (PMDs) in a single condition as well as differentially methylated regions (DMRs) between two conditions. We performed a rigorous benchmarking comparing existing approaches by evaluating the number, size, level of DNA methylation, boundaries, CpG content and coverage of the regions using several real datasets as well as the sensitivity and precision of the approaches using simulated data and show that MethyLasso performs best overall. MethyLasso is freely available at https://github.com/abardet/methylasso.

ABSTRACT. To fully understand gene regulation, it is necessary to have a thorough understanding of both the transcriptome and the enzymatic and RNA-binding activities that shape it. While many RNA-Seq-based tools have been developed to analyze the transcriptome, most only consider the abundance of sequencing reads along annotated patterns (such as genes). These annotations are typically incomplete, leading to errors in the differential expression analysis. To address this issue, we present DiffSegR - an R package that enables the discovery of transcriptome-wide expression differences between two biological conditions using RNA-Seq data. DiffSegR does not require prior annotation and uses a multiple changepoints detection algorithm to identify the boundaries of differentially expressed regions in the per-base log2 fold change. In a few minutes of computation, DiffSegR could rightfully predict the role of chloroplast ribonuclease Mini-III in rRNA maturation and chloroplast ribonuclease PNPase in (3′/5′)-degradation of rRNA, mRNA and tRNA precursors as well as intron accumulation. We believe DiffSegR will benefit biologists working on transcriptomics as it allows access to information from a layer of the transcriptome overlooked by the classical differential expression analysis pipelines widely used today. DiffSegR is available at https://aliehrmann.github.io/DiffSegR/index.html.

ABSTRACT. Short Open Reading Frames (sORFs) are ubiquitous genomic elements that have been overlooked for years, essentially due to their short length (< 100 residues) and the use of alternative start codons (other than AUG). However, some may encode functional peptides, so-called sORF-encoded peptides (sPEPs), the functions of which remain mainly unknown. In this study, we propose a system-level approach to determine the functions of sPEPs in monocytes. We first predicted the interactions of sPEPs with canonical proteins and analyzed the interfaces of interactions as well as the set of canonical proteins interacting with sPEPs. Second, by joining these sPEP-canonical proteins interactions with the human interactome, we predicted the first sPEP interactome network to date. Based on its topology, we then predicted the function of the sPEPs. Our results suggest that the majority of sPEPs are involved in key biological functions, including regulatory functions, metabolism, and signaling. Overall, the diversity in the predicted functions of the sPEPs underline the prevalence of their role in different biological mechanisms, suggesting that they are major regulatory actors.