Section: New Results

Bioinformatics Analysis

Study of marine plankton holobionts

Participants : Camille Marchet, Pierre Peterlongo.

We derived from the quasi-dictionary (described in previous section) a tool called Short Read Connector (SRC), able to find pairs of similar reads intra or inter read sets. We used SRC in meta-transcriptomics context to identify the actors of a symbiosis and help the assembly [44], [31]. The framework is the study of marina holobionts (host and its community of symbionts) for which few is known about the actors. In order to retrieve the functions that characterize such holobionts, RNA-seq reads from the sequencing of the whole holobiont are assembled de novo. Such assembly is prone to produce chimeras. Thus SRC is used to index sequences (reads, EST, assembled genes...) known to be close to the host and symbionts of the holobiont. Then, thanks to SRC's ability to find similarity between sequences even at a large scale, by querying reads of the holobiont we identify those similar to the host or symbionts. We report four categories: host, symbiont, shared and unassigned that can be assembled in a parallel way. As a first step we validate the SRC+assembly approach by comparing our result to literature with two known holobionts with eukaryote hosts (Orbicella faveolata, Xestospongia muta). We show that our approach can compare to previous results. In a second step we lean on a protist (Collodaria) holobiont for which the actors are poorly known. No assembled sequences exist in the literature so we compare the pipeline SRC+assembly to a sole assembly pipeline. Our main achievement is to highly reduce (up to $\sim$40%) the number of chimeras in the assembly compared to the sole assembly pipeline.

Pea aphid metagenomics

Participants : Cervin Guyomar, Fabrice Legeai, Claire Lemaitre.

We worked on a framework adapted to the study of genomic diversity and evolutionary dynamics of the pea aphid symbiotic community from an extensive set of metagenomics datasets. The framework is based on mapping to reference genomes and whole genome SNP-calling. We explored the genotypic diversity associated to the different symbionts of the pea aphid at several scales : across host biotypes, amongst individuals of the same biotype, or within individual aphids. Thorough phylogenomic analyses highlighted that the evolutionary dynamics of symbiotic associations strongly varied depending on the symbiont, reflecting different histories and possible constraints [40], [30].

Assembly and comparison of two genomes of highly polyphagous lepidopteran pests

Participants : Fabrice Legeai, Claire Lemaitre.

In this study, two genomes of an agronomical important lepidopteran pest, the noctuid moth Spodoptera frugiperda, were sequenced and compared, giving significant insights to the mechanisms involved in host-plant adaptation and speciation of this organism. In particular, we described the large expansion of gustatory receptors and detoxification genes among this polyphagous pest compared to other specialist Lepidoptera, and emphasizes the role of these 2 gene families in the evolution of one of the world’s worst agricultural pests. We also provided the genome assemblies, gene annotations and whole genome alignments of both strains, and the comparison of both to a reference moth genome (Bombyx mori). For these purposes, several original methods were developed i) to correct genome assembly errors due to the high level of heterozygosity and ii) to extract structural variant calls from whole genome alignments [15].

Benchmark of de novo read dataset compression tools

Participants : Gaetan Benoit, Dominique Lavenier, Claire Lemaitre.

In this book chapter, we review the different approaches and their tools developed so far to compress sequencing data files. We detail the algorithms for each of the three main types of data contained in such files for each read : the header, the DNA and the quality sequences. We also provide a thorough benchmark of the numerous available tools on various sequencing datasets, evaluating the compression ratio as well as the running time and memory usage performances [33].

Genomics of the agro-ecosystems pests

Participants : Fabrice Legeai, Claire Lemaitre.

Within a large international network of biologists, GenScale has contributed to various projects for identifying important components such as protein coding or non coding genes involved in the adaptation of major agricultural pests to their environment. We provided or participated to the assembly and the annotation of 4 new aphids [17], [22], and 5 parasitic wasps. Following specific agreement or policy, these new genomes and annotations are available for a restricted consortium or a large community through the BioInformatics platform for Agro-ecosystems Arthropods (http://bipaa.genouest.org/is). Moreover our engagement in the agronomical pest genomics led to our contribution to other projects such as epigenetics and chromatin structure analysis [18], or the analysis of population genetics data for identifying hotspots of selection in the nematode Globodera pallida genome [14].

Comparison of approaches for finding alternative splicing events in RNA-seq

Participant : Camille Marchet.

In this work we compared an assembly-first and a mapping-first approach to analyze RNA-seq data and find alternative splicing (AS) events. Assembly-first approach enables to identify novel AS events and to detect events in paralog genes that are hard to find using mapping because of the multi-mapping results. On the other hand, the mapping-first approach is more sensitive and detects AS events in lowly expressed genes, and is also able to find AS events with exons containing transposable elements. In addition we support these results with experimental validation. We showed that in order to extensively study the alternative splicing via RNA-seq data and retrieve the most candidates, both approaches should be led. We provide a pipeline consituted of parallel local de novo assembly executed by KisSplice and mapping using a novel mapping workflow called FaRLine [37].

Microbial communities interaction between plant and their bioagressors

Participants : Susete Alves Carvalho, Fabrice Legeai, Claire Lemaitre, Pierre Peterlongo, Dominique Lavenier.

GenScale actively collaborates with the INRA group ‘plant-microbial communities interactions’ (IGEPP, Rennes) that analyze the interaction between plant, their associated microbial communities and different bioagressors. The ambition of the project is to understand the link between the taxonomic biodiversity of the microbiota and their functional diversity in relation with plant physiology and plant-bioagressors interactions. For this last point, an integrated metatranscritomic approach is developped. Beside wet lab and sequence productions, bioinformatics tools are needed and meta-transcriptomic pipelines analysis arecurrently developped based on the GenScale expertise.