RAPiD: a rapid and accurate plant pathogen identification pipeline for on-site nanopore sequencing

Abstract

Nanopore sequencing technology has enabled the rapid, on-site taxonomic identification of samples from anything and anywhere. However, sequencing errors, inadequate databases, as well as the need for bioinformatic expertise and powerful computing resources, have hampered the widespread use of the technology for pathogen identification in the agricultural sector. Here we present RAPiD, a lightweight and accurate real-time taxonomic profiling pipeline. Compared to other metagenomic profilers, RAPiD had a higher classification precision achieved through the use of a curated, non-redundant database of common agricultural pathogens and extensive quality filtering of alignments. On a fungal, bacterial and mixed mock community RAPiD was the only pipeline to detect all members of the communities. We also present a protocol for in-field sample processing enabling pathogen identification from plant sample to sequence within 3 h using low-cost equipment. With sequencing costs continuing to decrease and more high-quality reference genomes becoming available, nanopore sequencing provides a viable method for rapid and accurate pathogen identification in the field. A web implementation of the RAPiD pipeline for real-time analysis is available at https://agrifuture.senckenberg.de.

Keywords: Metagenomics; Nanopore sequencing; On-site sequencing; Plant pathogens; Realtime genomics.

Figure 1. Workflow of the RAPiD pipeline.

Workfow of the RAPiD pipeline, divided into data input (blue), read processing, index creation and read alignment (orange), and report generation (green). The web-application enables additional basecalling of raw fast5 files and extension of the reference database (pathogen index) through submission of new genome assemblies.

Figure 2. Overview of in-field plant pathogen detection with necessary equipment.

Pictures of infected plant samples with normalised alignment score and mean per base identity for sequence matches: (A) common wheat plant with wheat leaf rust (Puccinia triticina); (B) tomato plant with leaf blight (Alternaria spp.); (C) apple plant leaf with apple scab (Venturia inaequalis); (D) instrument set-up for DNA extraction and sequencing in the field on an Oxford Nanopore Technology Mk1C device.

Figure 3. RAPiD pipeline values for alignment metrics against the reference database.

Values for alignment metrics of target (grey) and non-target species (black) against a reference database containing representative genomes of each target species. Cut-off values excluding over 99% of false positive matches are shown as red lines in selected metrics.