Abstract
To evaluate new drugs, the immune system should be considered. Here we evaluated a proof-of-concept that uncovers bacterial-leukocyte interactions. Analyzing longitudinal leukocyte data from bovines infected with either methicillin-resistant (MRSA) or methicillin-susceptible (MSSA) Staphylococcus aureus, two methods were investigated: (i) an approach that assesses lymphocytes, monocytes, or neutrophils, separately, and (ii) a method that, using dimensionless indicators (products, ratios, or combinations derived from leukocyte data), explores the dynamics of leukocyte relationships in three-dimensional (3D) space and identifies data subsets of informative value.
The classic approach not always distinguished infected from non-infected cows. In contrast, the alternative approach differentiated noninfected from infected animals and distinguished early MRSA from early MSSA and late MRSA infections.
Discrimination was associated with the use of dimensionless indicators. When measured in 3D space, such indicators generated a very large number of combinations, which helped detect data subsets usually unobserved, such as non-overlapping infection-negative and -positive subsets, and several disease stages. The validity of such data subsets was determined with biologically interpretable data.
This graphic, pattern recognition-based information system included but did not depend on any one number or variable. Because it can detect functions (relationships that involve two or more elements), in real time, if shown reproducible, the analysis of complex hostmicrobial dynamics could be used to evaluate antimicrobials.
Keywords: Infection, MRSA, MSSA, complexity, systems, dynamics, three-dimensional, cutoff-free discrimination.