Interpretation of microbiota-based diagnostics by explaining individual classifier decisions.

BACKGROUND: The human microbiota is associated with various disease states and holds a great promise for non-invasive diagnostics. However, microbiota data is challenging for traditional diagnostic approaches: It is high-dimensional, sparse and comprises of high inter-personal variation. State of the art machine learning tools are therefore needed to achieve this goal. While these tools have the ability to learn from complex data and interpret patterns therein that cannot be identified by humans, they often operate as black boxes, offering no insight into their decision-making process. In most cases, it is difficult to represent the learning of a classifier in a comprehensible way, which makes them prone to be mistrusted, or even misused, in a clinical environment. In this study, we aim to elucidate microbiota-based classifier decisions in a biologically meaningful context to allow their interpretation. RESULTS: We applied a method for explanation of classifier decisions on two microbiota datasets of increasing complexity: gut versus skin microbiota samples, and inflammatory bowel disease versus healthy gut microbiota samples. The algorithm simulates bacterial species as being unknown to a pre-trained classifier, and measures its effect on the outcome. Consequently, each patient is assigned a unique quantitative estimation of which species in their microbiota defined the classification of their sample. The algorithm was able to explain the classifier decisions well, demonstrated by our validation method, and the explanations were biologically consistent with recent microbiota findings. CONCLUSIONS: Application of a method for explaining individual classifier decisions for complex microbiota analysis proved feasible and opens perspectives on personalized therapy. Providing an explanation to support a microbiota-based diagnosis could guide decisions of clinical microbiologists, and has the potential to increase their confidence in the outcome of such decision support systems. This may facilitate the development of new diagnostic applications.

MEDI4893* Promotes Survival and Extends the Antibiotic Treatment Window in a Staphylococcus aureus Immunocompromised Pneumonia Model.

Immunocompromised individuals are at increased risk of Staphylococcus aureus pneumonia. Neutralization of alpha-toxin (AT) with the monoclonal antibody (MAb) MEDI4893* protects normal mice from S. aureus pneumonia; however, the effects of the MAb in immunocompromised mice have not been reported. In this study, passive immunization with MEDI4893* increased survival rates and reduced bacterial numbers in the lungs in an immunocompromised murine S. aureus pneumonia model. Lungs from infected mice exhibited alveolar epithelial damage, protein leakage, and bacterial overgrowth, whereas lungs from mice passively immunized with MEDI4893* retained a healthy architecture, with an intact epithelial barrier. Adjunctive therapy or prophylaxis with a subtherapeutic MEDI4893* dose combined with subtherapeutic doses of vancomycin or linezolid improved survival rates, compared with the monotherapies. Furthermore, coadministration of MEDI4893* with vancomycin or linezolid extended the antibiotic treatment window. These data suggest that MAb-mediated neutralization of AT holds promise in strategies for prevention and adjunctive therapy among immunocompromised patients.

Frequency and peak stretch magnitude affect alveolar epithelial permeability.

The present study measured stretch-induced changes in transepithelial permeability to uncharged tracers (1.5-5.5 A) using cultured monolayers of alveolar epithelial type-I like cells. Cultured alveolar epithelial cells were subjected to uniform cyclic (0, 0.25 and 1.0 Hz) biaxial stretch from 0% to 12, 25 or 37% change in surface area (DeltaSA) for 1 h. Significant changes in permeability of cell monolayers were observed when stretched from 0% to 37% DeltaSA at all frequencies, and from 0% to 25% DeltaSA only at high frequency (1 Hz), but not at all when stretched from 0% to 12% DeltaSA compared with unstretched controls. At stretch oscillation amplitudes of 25 and 37% DeltaSA, imposed at 1 Hz, tracer permeability increased compared with that at 0.25 Hz. Cells subjected to a single stretch cycle at 37% DeltaSA (0.25 Hz), to simulate a deep sigh, were not distinguishable from unstretched controls. Reducing stretch oscillation amplitude while maintaining a peak stretch of 37% DeltaSA (0.25 Hz) via the application of a simulated post-end-expiratory pressure did not protect barrier properties. In conclusion, peak stretch magnitude and stretch frequency were the primary determining factors for epithelial barrier dysfunction, as opposed to oscillation amplitude.

Deformation of the human brain induced by mild acceleration.

Rapid deformation of brain matter caused by skull acceleration is most likely the cause of concussion, as well as more severe traumatic brain injury (TBI). The inability to measure deformation directly has led to disagreement and confusion about the biomechanics of concussion and TBI. In the present study, brain deformation in human volunteers was measured directly during mild, but rapid, deceleration of the head (20-30 m/sec2 peak, approximately 40 msec duration), using an imaging technique originally developed to measure cardiac deformation. Magnetic resonance image sequences with imposed "tag" lines were obtained at high frame rates by repeating the deceleration and acquiring a subset of image data each repetition. Displacements of points on tag lines were used to estimate the Lagrangian strain tensor field. Qualitative (visual) and quantitative (strain) results illustrate clearly the deformation of brain matter due to occipital deceleration. Strains of 0.02-0.05 were typical during these events (0.05 strain corresponds roughly to a 5% change in the dimension of a local tissue element). Notably, compression in frontal regions and stretching in posterior regions were observed. The motion of the brain appears constrained by structures at the frontal base of the skull; it must pull away from such constraints before it can compress against the occipital bone. This mechanism is consistent with observations of contrecoup injury in occipital impact.