SeqScreen-Nano: a computational platform for rapid, in-field characterization of previously unseen pathogens

Abstract

The COVID-19 pandemic forever underscored the need for biosurveillance platforms capable of rapid detection of previously unseen pathogens. Oxford Nanopore Technology (ONT) couples long-read sequencing with in-field capability, opening the door to real-time, in-field biosurveillance. Though a promising technology, streaming assignment of accurate functional and taxonomic labels with nanopore reads remains challenging given: (i) individual reads can span multiple genes, (ii) individual reads may contain truncated genes, and pseudogenes, (iii) the error rate of the ONT platform that may introduce frameshifts and missense errors, and (iv) the computational costs of read-by-read analysis may exceed that of in-field computational equipment. Altogether, these challenges highlight a need for novel computational approaches. To this end, we describe SeqSeqscreen-Nano, a novel and portable computational platform for the characterization of novel pathogens. Based on results from simulated and synthetic microbial communities, SeqScreen-Nano can identify Open Reading Frames (ORFs) across the length of raw ONT reads and then use the predicted ORFs for accurate functional characterization and taxonomic classification. SeqScreen-Nano can run efficiently in a memory-constrained environment (less than 32GB of RAM), allowing it to be utilized in resource-limited settings. SeqScreen-Nano can also process reads directly from the ONT MinION sequencing device, enabling rapid, in-field characterization of previously unseen pathogens. SeqScreen-Nano (v4.0) is available on GitLab at:https://gitlab.com/treangenlab/seqscreen

Validation of an unbiased metagenomic detection assay for RNA viruses in viral transport media and plasma

“…Untargeted sequence analysis was performed using Centrifuge (Kim et al 2016) with a customized virus database to identify putative NCBI taxonomy identifiers (taxids) and input those taxids into the 'reference inference module' of SeqScreen-Nano (Balaji et al 2023b(Balaji et al , 2023a.…”

“…Mapping metrics were re-calculated and further statistical analysis was performed to determine the likelihood of presence or absence of individual genomes based on a rubric. The SeqScreen-Nano 'reference inference module' analysis, metrics, and rubrics are originally described in Balaji et al (2023a). Modification to the threshold criteria was made to be more inclusive of the reference genomes taken from the first mapping of individual reference genomes to the second mapping of a concatenated reference.…”

“…Modification to the threshold criteria was made to be more inclusive of the reference genomes taken from the first mapping of individual reference genomes to the second mapping of a concatenated reference. This included reducing the original calculated in SeqScreen-Nano 'coverage score' threshold, which is the ratio of the observed 'breadth of coverage' and 'expected breadth of coverage' for each taxon in a sample (Balaji et al 2023a), from a minimum of 0.70 to 0.10. The large ratio of 0.7 for the 'coverage score' was not ideal for smaller genomes for viruses, which had inflated expected coverages compared to genomes sizes of bacteria.…”

Lightweight taxonomic profiling of long-read sequenced metagenomes with Lemur and Magnet

“…Magnet calculates the observed breadth and depth of coverage for the alignment with samtools coverage for two cases: (1) including all alignments with MAPQ ≥ 1 and (2) including only primary alignments with MAPQ ≥ 20. For both cases, Magnet calculates the expected breadth of coverage based on the abundance and the coverage score, defined as the ratio of the observed breadth of coverage and expected breadth of coverage [4]. The coverage score is used to measure uniformity of the alignment distribution along the reference genome.…”