Significance Assessing the threat posed by bacterial samples is fundamentally important to safeguarding human health. Whole-genome sequence analysis of bacteria provides a route to achieving this goal. However, this approach is fundamentally constrained by the scope, the diversity, and our understanding of the bacterial genome sequences that are available for devising threat assessment schemes. For example, genome-based strategies offer limited utility for assessing the threat associated with pathogens that exploit novel virulence mechanisms or are recently emergent. To address these limitations, we developed PathEngine, a machine learning strategy that features the use of phenotypic hallmarks of pathogenesis to assess pathogenic threat. PathEngine successfully classified potential pathogenic threats with high accuracy and thereby establishes a phenotype-based, sequence-independent pipeline for threat assessment.

Limited data significantly hinders our capability of biothreat assessment of novel bacterial strains. Integration of data from additional sources that can provide context about the strain can address this challenge. Datasets from different sources, however, are generated with a specific objective and which makes integration challenging. Here, we developed a deep learning-based approach called the neural network embedding model (NNEM) that integrates data from conventional assays designed to classify species with new assays that interrogate hallmarks of pathogenicity for biothreat assessment. We used a dataset of metabolic characteristics from a de-identified set of known bacterial strains that the Special Bacteriology Reference Laboratory (SBRL) of the Centers for Disease Control and Prevention (CDC) has curated for use in species identification. The NNEM transformed results from SBRL assays into vectors to supplement unrelated pathogenicity assays from de-identified microbes. The enrichment resulted in a significant improvement in accuracy of 9% for biothreat. Importantly, the dataset used in our analysis is large, but noisy. Therefore, the performance of our system is expected to improve as additional types of pathogenicity assays are developed and deployed. The proposed NNEM strategy thus provides a generalizable framework for enrichment of datasets with previously collected assays indicative of species.