We describe DAMIAN, an open source bioinformatics tool designed for the identification of pathogenic microorganisms in diagnostic samples. By using authentic clinical samples and comparing our results to those from established analysis pipelines as well as conventional diagnostics, we demonstrate that DAMIAN rapidly identifies pathogens in different diagnostic entities, and accurately classifies viral agents down to the strain level. We furthermore show that DAMIAN is able to assemble full-length viral genomes even in samples co-infected with multiple virus strains, an ability which is of considerable advantage for the investigation of outbreak scenarios. While DAMIAN, similar to other pipelines, analyzes single samples to perform classification of sequences according to their likely taxonomic origin, it also includes a tool for cohort-based analysis. This tool uses cross-sample comparisons to identify sequence signatures that are frequently present in a sample group of interest (e.g., a disease-associated cohort), but occur less frequently in control cohorts. As this approach does not require homology searches in databases, it principally allows the identification of not only known, but also completely novel pathogens. Using samples from a meningitis outbreak, we demonstrate the feasibility of this approach in identifying enterovirus as the causative agent.