With recent advances in natural language processing, rationalization becomes an essential self-explaining diagram to disentangle the black box by selecting a subset of input texts to account for the major variation in prediction. Yet, existing association-based approaches on rationalization cannot identify true rationales when two or more snippets are highly inter-correlated and thus provide a similar contribution to prediction accuracy, so-called spuriousness. To address this limitation, we novelly leverage two causal desiderata, non-spuriousness and efficiency, into rationalization from the causal inference perspective. We formally define a series of probabilities of causation based on a newly proposed structural causal model of rationalization, with its theoretical identification established as the main component of learning necessary and sufficient rationales. The superior performance of the proposed causal rationalization is demonstrated on real-world review and medical datasets with extensive experiments compared to state-of-the-art methods.

Heterogeneity and comorbidity are two interwoven challenges associated with various healthcare problems that greatly hampered research on developing effective treatment and understanding of the underlying neurobiological mechanism. Very few studies have been conducted to investigate heterogeneous causal effects (HCEs) in graphical contexts due to the lack of statistical methods. To characterize this heterogeneity, we first conceptualize heterogeneous causal graphs (HCGs) by generalizing the causal graphical model with confounder-based interactions and multiple mediators. Such confounders with an interaction with the treatment are known as moderators. This allows us to flexibly produce HCGs given different moderators and explicitly characterize HCEs from the treatment or potential mediators on the outcome. We establish the theoretical forms of HCEs and derive their properties at the individual level in both linear and nonlinear models. An interactive structural learning is developed to estimate the complex HCGs and HCEs with confidence intervals provided. Our method is empirically justified by extensive simulations and its practical usefulness is illustrated by exploring causality among psychiatric disorders for trauma survivors. Code implementing the proposed algorithm is open-source and publicly available at: https://github.com/richard-watson/ISL.

Phytophthora ramorum is an invasive, broad host range pathogen that causes ramorum blight and sudden oak death in forest landscapes of western North America. In commercial nurseries, asymptomatic infections of nursery stock by P. ramorum and other Phytophthora species create unacceptable risk and complicate inspection and certification programs designed to prevent introduction and spread of these pathogens. In this study, we continue development of a volatile organic compound (VOC)-based test for detecting asymptomatic infections of P. ramorum in Rhododendron sp. We confirmed detection of P. ramorum from volatiles collected from asymptomatic root-inoculated Rhododendron plants in a nursery setting, finding that the VOC profile of infected plants is detectably different from that of healthy plants, when measured from both ambient VOC emissions and VOCs extracted from leaf material. Predicting infection status was successful from ambient volatiles, which had a mean area under the curve (AUC) value of 0.71 ± 0.17, derived from corresponding receiver operating characteristic curves from an extreme gradient boosting discriminant analysis. This finding compares with that of extracted leaf volatiles, which resulted in a lower AUC value of 0.51 ± 0.21. In a growth chamber, we contrasted volatile profiles of asymptomatic Rhododendron plants having roots infected with one of three pathogens: P. ramorum, P. cactorum, and Rhizoctonia solani. Each pathogen induced unique and measurable changes, but generally the infections reduced volatile emissions until 17 weeks after inoculation, when emissions trended upward relative to those of mock-inoculated controls. Forty-five compounds had significant differences compared with mock-inoculated controls in at least one host-pathogen combination.

Respiratory viral infections are considered a major public health threat, and breath metabolomics can provide new ways to detect and understand how specific viruses affect the human pulmonary system. In this pilot study, we characterized the metabolic composition of human breath for an early diagnosis and differentiation of influenza viral infection, as well as other types of upper respiratory viral infections. We first studied the non-specific effects of planned seasonal influenza vaccines on breath metabolites in healthy subjects after receiving the immunization. We then investigated changes in breath content from hospitalized patients with flu-like symptoms and confirmed upper respiratory viral infection. The exhaled breath was sampled using a custom-made breath condenser, and exhaled breath condensate (EBC) samples were analysed using liquid chromatography coupled to quadruplole-time-of-flight mass spectrometer (LC-qTOF). All metabolomic data was analysed using both targeted and untargeted approaches to detect specific known biomarkers from inflammatory and oxidative stress biomarkers, as well as new molecules associated with specific infections. We were able to find clear differences between breath samples collected before and after flu vaccine administration, together with potential biomarkers that are related to inflammatory processes and oxidative stress. Moreover, we were also able to discriminate samples from patients with flu-related symptoms that were diagnosed with confirmatory respiratory viral panels (RVPs). RVP positive and negative differences were identified, as well as differences between specific viruses defined. These results provide very promising information for the further study of the effect of influenza A and other viruses in human systems by using a simple and non-invasive specimen like breath.

An automated microfluidic sample preparation multiplexer (SPM) has been developed and evaluated for Ebola virus detection. Metered air bubbles controlled by microvalves are used to improve bead-solution mixing thereby enhancing the hybridization of the target Ebola virus RNA with capture probes bound to the beads. The method uses thermally stable 4-formyl benzamide functionalized (4FB) magnetic beads rather than streptavidin coated beads with a high density of capture probes to improve the target capture efficiency. Exploiting an on-chip concentration protocol in the SPM and the single molecule detection capability of the antiresonant reflecting optical waveguide (ARROW) biosensor chip, a detection limit of 0.021pfu/mL for clinical samples is achieved without target amplification. This RNA target capture efficiency is two orders of magnitude higher than previous results using streptavidin beads and the limit of detection (LOD) improves 10×. The wide dynamic range of this technique covers the whole clinically applicable concentration range. In addition, the current sample preparation time is ~1h which is eight times faster than previous work. This multiplexed, miniaturized sample preparation microdevice establishes a key technology that intended to develop next generation point-of-care (POC) detection system.

To evaluate associations of premature ovarian failure (POF) with mortality and morbidity in Asian populations.We identified 1,003 cases of POF among 36,402 postmenopausal women who participated in the Shanghai Women's Health Study, a population-based cohor

We present the activities of the 'New Physics' working group for the 'Physics at TeV Colliders' workshop (Les Houches, France, 1-19 June, 2015). Our report includes new physics studies connected with the Higgs boson and its properties, direct search strategies, reinterpretation of the LHC results in the building of viable models and new computational tool developments. Important signatures for searches for natural new physics at the LHC and new assessments of the interplay between direct dark matter searches and the LHC are also considered.

We report a measurement of the branching fraction for ψ(3770)→γχc1 and search for the transition ψ(3770)→γχc2 based on 2.92 fb-1 of e+e- data accumulated at s=3.773 GeV with the BESIII detector at the BEPCII collider. We measure B(ψ(3770)→γχc1)=(2.48±0.15±0.23)×10-3, which is the most precise measurement to date. The upper limit on the branching fraction of ψ(3770)→γχc2 at a 90% confidence level is B(ψ(3770)→γχc2)<0.64×10-3. The corresponding partial widths are Γ(ψ(3770)→γχc1)=(67.5±4.1±6.7) keV and Γ(ψ(3770)→γχc2)<17.4 keV.

The process e+e−→ pp̄π0has been studied by analyzing data collected at (formula presented) GeV, at (formula presented) GeV, and during a ψ(3770) line shape scan with the BESIII detector at the BEPCII collider. The Born cross section of p̄pπ0in the vicinity of the ψ(3770) is measured, and the Born cross section of ψ(3770) → p̄pπ0is extracted considering interference between resonant and continuum production amplitudes. Two solutions with the same probability and a significance of 1.5 σ are found. The solutions for the Born cross section of ψ(3770) → p̄pπ0are 33.8 ± 1.8 ± 2.1 pb and (formula presented) pb (< 0.22 pb at a 90% confidence level). Using the estimated cross section and a constant decay amplitude approximation, the cross section σ(p̄p → ψ(3770)π0) is calculated for the kinematic situation of the planned P̄ANDA experiment. The maximum cross section corresponding to the two solutions is expected to be less than 0.79 nb at 90% confidence level and 122 ± 10 nb at a center-of-mass energy of 5.26 GeV.