OHEJP-RaDAR-D-JRP3-1.4 Test and parameterization of the assembly pipeline for metagenomics data

Context

In parallel to the development of a complete database of plasmids [1] (see D1.5), we successfully performed plasmid assembly, typing and/or mobility prediction from sequencing short-reads of bacterial genomes [2] based on PlasmidSpades assembly [3] and/or ARTwork genomic assembly [4] coupled with Abricate [5] and PlasmidFinder database [6] or MOB-suite tools (i.e. MOB-recon and MOB-typer) [7] (see D1.3).

Objectives

We plan to (i) setup a metagenomics workflow to perform de novo meta-assembly from sequencing paired-end short reads of meta-samples, and to (ii) integrate our complete database of plasmids [1] in the MOB-suite tools [7] in order to (iii) type and predict mobility of plasmids from metadata.

Methods

The de novo meta-assembly is performed successively with Trimmomatic for trimming of reads [8], MetaSpades for assembly [9] and MetaQuast for quality checking [10] (i). The integration of our complete database of plasmids [1] in the MOB-suite tools [7] is done with the MOB-cluster tool. The plasmids from meta-assembly are typed and mobility predicted with our complete database of plasmids [1] and the MOB-suite tools [7].

Results

De novo meta-assembly of paired-end short reads from a sewage water meta-sample has been done in bash successfully with Trimmomatic [8] and MetaSpades [9] but the quality check with MetaQuast [10] showed a highly fragmented meta-assembly (i). The MOB-cluster tool allowed integration of our complete database of plasmids [1] in the MOB-suite tools [7] (ii). After decreasing of the meta-assembly fragmentation [10] based on less drastic parameters related to trimming [8] and de novo meta-assembly steps [9], we plan to apply the MOB-suite tools [7] integrating our complete database of plasmids [1] in order to screen types and mobility of plasmids from meta-assembly of paired-end short reads.