Probe-based metagenomics pathogen detection: an added value for complex diagnosis. Results of a pilot experiment at the Portuguese National Institute of Health (INSA)

This poster was presented on the 9th International Conference on Clinical Metagenomics (ICCMg) in Geneva, Switzerland. The authors Rita Ferreira, Luís Coelho, Daniel Sobral and João Dourado Santos contributed equally to the elaboration of this experiment. You can find an abstract for this poster below: Metagenomics Next-Generation Sequencing (mNGS) promises accurate and rapid detection of unidentified and/or uncultivable organisms to aid clinical diagnosis. However, mNGS presents several challenges, such as high costs, small microbial/human genetic material ratio, and demanding laboratorial and computational resources. Probe-based pathogen enrichment mNGS can strike a balance between hypothesis-free shotgun mNGS and conventional pathogen-specific detection. In this context, we tested two Illumina panels for probe-based pathogen enrichment and detection, the Respiratory and the Urinary Pathogen ID/AMR panels (RPIP and UPIP), which, together, target 383 pathogenic agents (virus, bacteria, fungi and parasites). The selections of samples intended to mimic samples’ heterogeneity of those routinely received at INSA: 99 clinical samples of different nature (e.g., cerebrospinal fluid, plasma, serum, urine, swabs, biopsies, etc.) for which at least one pathogen had been previously identified by (q)PCR were selected, with qPCR Cycle threshold (Ct) values ranging from 9.7 to 41.3. Overall, a total of 114 previously detected pathogens were targeted in this study. Illumina’s turnkey bioinformatics solution (Explify®, BaseSpace) yielded hits for 73.7% of the expected pathogens. Still, as the use of NGS in clinical diagnosis requires robust bioinformatics confirmation, we further applied our start-to-end bioinformatics pipeline (INSaFLU-TELEVIR+) for viral and bacterial metagenomics detection (an expanded version of INSaFLU-TELEVIR, https://insaflu.insa.pt/). We could obtain a combined detection rate of 79.8%, reaching as high as 89.7% for viruses and 65.7% for bacteria. Although higher Ct values were clearly a significant factor underlying miss detections, we could still detect most pathogens even at Cts>30, with 39.5 being the highest Ct value for a detected pathogen in our dataset. Despite the laborious pre-NGS wet-lab procedures, this pilot study suggest that the RPIP/UPIP probe-based Illumina panels, coupled with in-depth downstream bioinformatics, may constitute a powerful tool to help difficult diagnosis and support clinical decision.