Generating automated kidney transplant biopsy reports combining molecular measurements with ensembles of machine learning classifiers

We previously reported a system for assessing rejection in kidney transplant biopsies using microarray‐based gene expression data, the Molecular Microscope^® Diagnostic System (MMDx). The present study was designed to optimize the accuracy and stability of MMDx diagnoses by replacing single machine learning classifiers with ensembles of diverse classifier methods. We also examined the use of automated report sign‐outs and the agreement between multiple human interpreters of the molecular results. Ensembles generated diagnoses that were both more accurate than the best individual classifiers, and nearly as stable as the best, consistent with expectations from the machine learning literature. Human experts had ≈93% agreement (balanced accuracy) signing out the reports, and random forest‐based automated sign‐outs showed similar levels of agreement with the human experts (92% and 94% for predicting the expert MMDx sign‐outs for T cell–mediated (TCMR) and antibody‐mediated rejection (ABMR), respectively). In most cases disagreements, whether between experts or between experts and automated sign‐outs, were in biopsies near diagnostic thresholds. Considerable disagreement with histology persisted. The balanced accuracies of MMDx sign‐outs for histology diagnoses of TCMR and ABMR were 73% and 78%, respectively. Disagreement with histology is largely due to the known noise in histology assessments (ClinicalTrials.gov NCT01299168).     

Targeted Tissue and Cell-Free Tumor DNA Sequencing of Advanced Lung Squamous-Cell Carcinoma Reveals Clinically Significant Prevalence of Actionable Alterations

Background

Major guidelines do not recommend routine molecular profiling of lung squamous-cell carcinoma (LUSC) because the prevalence of actionable alterations is thought to be low. Increased utilization of next-generation sequencing (NGS), particularly with cell-free circulating tumor DNA, facilitates reevaluation of this premise.

Beta toxin catalyzes formation of nucleoprotein matrix in staphylococcal biofilms

Biofilms are surface-associated communities of microbes encompassed by an extracellular matrix. It is estimated that 80% of all bacterial infections involve biofilm formation, but the structure and regulation of biofilms are incompletely understood. Extracellular DNA (eDNA) is a major structural component in many biofilms of the pathogenic bacterium Staphylococcus aureus, but its role is enigmatic. Here, we demonstrate that beta toxin, a neutral sphingomyelinase and a virulence factor of S. aureus, forms covalent cross-links to itself in the presence of DNA (we refer to this as biofilm ligase activity, independent of sphingomyelinase activity) producing an insoluble nucleoprotein matrix in vitro. Furthermore, we show that beta toxin strongly stimulates biofilm formation in vivo as demonstrated by a role in causation of infectious endocarditis in a rabbit model. Together, these results suggest that beta toxin cross-linking in the presence of eDNA assists in forming the skeletal framework upon which staphylococcal biofilms are established.     

A minimally disruptive method for measuring water potential in planta using hydrogel nanoreporters

Leaf water potential is a critical indicator of plant water status, integrating soil moisture status, plant physiology, and environmental conditions. There are few tools for measuring plant water status (water potential) in situ, presenting a critical barrier for developing appropriate phenotyping (measurement) methods for crop development and modeling efforts aimed at understanding water transport in plants. Here, we present the development of an in situ, minimally disruptive hydrogel nanoreporter (AquaDust) for measuring leaf water potential. The gel matrix responds to changes in water potential in its local environment by swelling; the distance between covalently linked dyes changes with the reconfiguration of the polymer, leading to changes in the emission spectrum via Förster Resonance Energy Transfer (FRET). Upon infiltration into leaves, the nanoparticles localize within the apoplastic space in the mesophyll; they do not enter the cytoplasm or the xylem. We characterize the physical basis for AquaDust’s response and demonstrate its function in intact maize (Zea mays L.) leaves as a reporter of leaf water potential. We use AquaDust to measure gradients of water potential along intact, actively transpiring leaves as a function of water status; the localized nature of the reporters allows us to define a hydraulic model that distinguishes resistances inside and outside the xylem. We also present field measurements with AquaDust through a full diurnal cycle to confirm the robustness of the technique and of our model. We conclude that AquaDust offers potential opportunities for high-throughput field measurements and spatially resolved studies of water relations within plant tissues.

Plant life depends on water availability. In managing this demand, irrigated agriculture accounts for 70% of all human water use (1). Physiologically, the process of transpiration (E) dominates this demand for water (Fig. 1A): Solar thermal radiation and the unsaturated relative humidity in the atmosphere drive evaporation from the wet internal surfaces of leaves; this water loss pulls water up through the plant’s vascular tissue (xylem) and out of the soil. This flow occurs along a gradient in the chemical potential of water, or water potential, ψ [MPa] (2). Studies of water relations and stress physiology over the past decades have found that values of ψ along the path of E (the soil–plant–atmosphere continuum [SPAC]) correlate with plant growth, crop yield and quality, susceptibility to disease, and the balance between water loss due to E and the uptake and assimilation of carbon dioxide (water-use efficiency) (3–5).Open in a separate windowFig. 1.AquaDust as an in situ reporter of water potential (ψ). (A) Schematic representation of a maize plant undergoing transpiration (E) in a dynamic environment driven by solar thermal radiation (Qrad) and photosynthetically active radiation (PAR), wind speed (u), temperature (T), vapor pressure deficit (VPD), and soil water potential (ψsoil). Water flows through the plant (blue arrows) along a gradient in water potential (∇→ψ). Zones on the leaves infiltrated with AquaDust serve as reporters of the local leaf water potential, ψleaf, via a short (∼30 s), minimally invasive measurement of FRET efficiency (ζ) with a leaf clamp. (B) Schematic representations of infiltration of a suspension of AquaDust and of the distribution of AquaDust within the cross-section of a leaf. AquaDust passes through the stomata and localizes in the apoplastic spaces within the mesophyll; the particles are excluded from symplastic spaces and the vascular bundle. (C) Schematic diagrams showing mechanism of AquaDust response: The swollen, “wet” state when water potential in its local environment, ψenv=0 (i.e., no stress condition), results in low FRET between donor (green circles) and acceptor (yellow circles) dye (Upper); and the shrunken, “dry” state when ψenv<0 (i.e., stressed condition) results in high FRET between fluorophores, thereby altering the emission spectra (Lower). (D) Fluorescent dyes were chosen to minimize reabsorption of AquaDust emission from chlorophyll; comparison of representative fluorescent emission from AquaDust (donor peak at 520 nm and acceptor peak at 580 nm) with the absorption spectra of chlorophyll and autofluorescence of maize leaf.Due to the recognized importance of water potential in controlling plant function, plant scientists have spent considerable effort devising accurate and reliable methods to measure water potential of the soil, stem, and leaf (6). Of these, plant water potentials, and particularly leaf water potential (ψleaf), represent valuable indicators of plant water status because they integrate both environmental conditions (e.g., soil water availability and evaporative demand) and plant physiological processes (e.g., root water uptake, xylem transport, and stomatal regulation) (7, 8). To date, techniques to measure ψleaf remain either slow, destructive, or indirect. The current tools (e.g., Scholander pressure chamber, psychrometer, and pressure probe) involve disruption of the tissue, the microenvironment, or both (9–11). For example, the widely used pressure chamber requires excision of leaves or stems for the measurement of ψleaf. Other techniques, such as stem and leaf psychrometry, require intimate contact with the tissue, and accurate and repeatable measurements are difficult to obtain (9, 12). These limitations have hindered the study of spatiotemporal water-potential gradients along the SPAC and the development of high-throughput strategies to phenotype based on tissue water potential (13). Additionally, current methods for measuring ψleaf provide averages over tissues in the leaf. This characteristic makes the dissection of water relations on subleaf scales challenging, such that important questions remain, for example, about the partitioning of hydraulic resistances within leaves between the xylem and mesophyll (14–16).These outstanding challenges in the measurement of water status in planta motivated us to develop the measurement strategy presented here, AquaDust, with the following characteristics: 1) Minimally disruptive: Compatible with simple, rapid measurements on intact leaves. Fig. 1A presents our approach, in which AquaDust reporters infiltrated into the mesophyll of the leaf provide an externally accessible optical signal that correlates with the local water potential. 2) Localized: allowing for access to the values of water potential at a well-defined location along the path of transpiration in the leaf tissue. Fig. 1B shows a schematic representation of AquaDust particles localized in the apoplastic volume within the mesophyll, at the end of the hydraulic path for liquid water within the plant. 3) Sensitive and specific: capable of resolving water potentials across the physiologically relevant range (∼−3<ψ<0 MPa) and with minimal sensitivity to other physical (e.g., temperature) and chemical (e.g., pH) variables. Fig. 1C presents a schematic representation of an AquaDust particle formed of hydrogel, a highly tunable material that undergoes a structural response to changes in local water potential (swollen when wet; collapsed when dry). We couple the swelling behavior of the particle to an optical signal via the incorporation of fluorescence dyes (green and yellow circles in Fig. 1C) that undergo variable Förster Resonance Energy Transfer (FRET) as a function of spatial separation. Fig. 1D presents typical AquaDust spectra at high (wet; green curve) and low (dry; yellow curve) water potentials. A change in water potential leads to a change in the relative intensity of the two peaks in the AquaDust spectrum, such that the relative FRET efficiency, ζ=f(ID,IA), can serve as a measure of water potential. 4) Inert: nondisruptive of the physiological properties of the leaf (e.g., photosynthetic capacity, transpiration rate, etc.).In this paper, we present the development, characterization, and application of AquaDust. We show that AquaDust provides a robust, reproducible response of its fluorescence spectra to changes in leaf water potential in situ and across the usual physiological range. We apply our approach to quantify the spatial gradients of water potential along individual leaves undergoing active transpiration and across a range of soil water potentials. With these measurements, we show that the localization of AquaDust in the mesophyll allows us to quantify the importance of hydraulic resistances outside the xylem. We further use AquaDust to measure the diurnal dynamics of ψleaf under field conditions, with repeated measurements on individual, intact leaves. These measurements demonstrate the field-readiness of our techniques and validate the leaf hydraulic model we have developed. We conclude that AquaDust offers a powerful basis for tracking, spatially and temporally, water potential in planta to study the mechanisms by which it couples to both biological and physical processes to define plant function.     

Bronchoscopic management of airway obstruction in pediatric endobronchial tuberculosis

The present report describes a case of severe airway obstruction caused by endobronchial tuberculosis in an 11-year-old girl who was successfully treated by bronchoscopic balloon dilation. This case illustrates the insidious presentation and the increasingly important role of bronchoscopic intervention in the management of endobronchial tuberculosis. In addition, a brief literature review of the condition in the pediatric age group is included.