Diagnostic Accuracy of Shotgun Metagenomics for Bloodstream Infections Is Influenced by Bioinformatics Workflow Selection

Bloodstream infection (BSI) is a severe and often fatal condition, and a major cause of mortality in patients with hematological malignancies due to underlying conditions and anticancer therapy-induced immunodeficiency. Rapid identification of the causative pathogens is essential as BSI results in worsened prognosis, extended hospitalization, delays or dose reductions in therapy, and may progress to sepsis and septic shock if untreated. Shotgun metagenomics is a culture-independent technique capable of detecting a wide range of fungal, viral, and bacterial organisms along with their antimicrobial resistance genes. Several studies showed that shotgun metagenomics enables the diagnosis of BSI, specifically in cases where conventional methods/culture-dependent techniques fail to identify the causative pathogens. However, evaluation of the accuracy of the applied bioinformatics pipelines remains incomplete. This study aimed to compare and optimize four commonly used bioinformatics pipelines (BLAST, Kraken, Metaphlan, RTG Core) for shotgun metagenomics by assessing their accuracy in identifying pathogens in blood samples from patients with hematological malignancies and suspected BSI, with blood culture serving as the reference standard. Our work shows that the selection of bioinformatics pipelines for diagnosing BSI strongly affects the precision of the findings, and an optimized BLAST pipeline was superior to the alternatives, as it was the only method that accurately identified the causative pathogens.