timsTOF HT Improves Protein Identification and Quantitative Reproducibility for Deep Unbiased Plasma Protein Biomarker Discovery.

Mass spectrometry (MS) is a valuable tool for plasma proteome profiling and disease biomarker discovery. However, wide-ranging plasma protein concentrations, along with technical and biological variabilities, present significant challenges for deep and reproducible protein quantitation. Here, we evaluated the qualitative and quantitative performance of timsTOF HT and timsTOF Pro 2 mass spectrometers for analysis of neat plasma samples (unfractionated) and plasma samples processed using the Proteograph Product Suite (Proteograph) that enables robust deep proteomics sampling prior to mass spectrometry. Samples were evaluated across a wide range of peptide loading masses and liquid chromatography (LC) gradients. We observed up to a 76% increase in total plasma peptide precursors identified and a >2-fold boost in quantifiable plasma peptide precursors (CV < 20%) with timsTOF HT compared to Pro 2. Additionally, approximately 4.5 fold more plasma peptide precursors were detected by both timsTOF HT and timsTOF Pro 2 in the Proteograph analyzed plasma vs neat plasma. In an exploratory analysis of 20 late-stage lung cancer and 20 control plasma samples with the Proteograph, which were expected to exhibit distinct proteomes, an approximate 50% increase in total and statistically significant plasma peptide precursors (q < 0.05) was observed with timsTOF HT compared to Pro 2. Our data demonstrate the superior performance of timsTOF HT for identifying and quantifying differences between biologically diverse samples, allowing for improved disease biomarker discovery in large cohort studies. Moreover, researchers can leverage data sets from this study to optimize their liquid chromatography-mass spectrometry (LC-MS) workflows for plasma protein profiling and biomarker discovery. (ProteomeXchange identifier: PXD047854 and PXD047839).

Applications, promises, and pitfalls of deep learning for fluorescence image reconstruction.

Deep learning is becoming an increasingly important tool for image reconstruction in fluorescence microscopy. We review state-of-the-art applications such as image restoration and super-resolution imaging, and discuss how the latest deep learning research could be applied to other image reconstruction tasks. Despite its successes, deep learning also poses substantial challenges and has limits. We discuss key questions, including how to obtain training data, whether discovery of unknown structures is possible, and the danger of inferring unsubstantiated image details.

Minimizing bias in massive multi-arm observational studies with BCAUS: balancing covariates automatically using supervision.

BACKGROUND: Observational studies are increasingly being used to provide supplementary evidence in addition to Randomized Control Trials (RCTs) because they provide a scale and diversity of participants and outcomes that would be infeasible in an RCT. Additionally, they more closely reflect the settings in which the studied interventions will be applied in the future. Well-established propensity-score-based methods exist to overcome the challenges of working with observational data to estimate causal effects. These methods also provide quality assurance diagnostics to evaluate the degree to which bias has been removed and the estimates can be trusted. In large medical datasets it is common to find the same underlying health condition being treated with a variety of distinct drugs or drug combinations. Conventional methods require a manual iterative workflow, making them scale poorly to studies with many intervention arms. In such situations, automated causal inference methods that are compatible with traditional propensity-score-based workflows are highly desirable. METHODS: We introduce an automated causal inference method BCAUS, that features a deep-neural-network-based propensity model that is trained with a loss which penalizes both the incorrect prediction of the assigned treatment as well as the degree of imbalance between the inverse probability weighted covariates. The network is trained end-to-end by dynamically adjusting the loss term for each training batch such that the relative contributions from the two loss components are held fixed. Trained BCAUS models can be used in conjunction with traditional propensity-score-based methods to estimate causal treatment effects. RESULTS: We tested BCAUS on the semi-synthetic Infant Health & Development Program dataset with a single intervention arm, and a real-world observational study of diabetes interventions with over 100,000 individuals spread across more than a hundred intervention arms. When compared against other recently proposed automated causal inference methods, BCAUS had competitive accuracy for estimating synthetic treatment effects and provided highly concordant estimates on the real-world dataset but was an order-of-magnitude faster. CONCLUSIONS: BCAUS is directly compatible with trusted protocols to estimate treatment effects and diagnose the quality of those estimates, while making the established approaches automatically scalable to an arbitrary number of simultaneous intervention arms without any need for manual iteration.

Causal deep learning reveals the comparative effectiveness of antihyperglycemic treatments in poorly controlled diabetes.

Type-2 diabetes is associated with severe health outcomes, the effects of which are responsible for approximately 1/4th of the total healthcare spending in the United States (US). Current treatment guidelines endorse a massive number of potential anti-hyperglycemic treatment options in various combinations. Strategies for optimizing treatment selection are lacking. Real-world data from a nationwide population of over one million high-risk diabetic patients (HbA1c ≥ 9%) in the US is analyzed to evaluate the comparative effectiveness for HbA1c reduction in this population of more than 80 different treatment strategies ranging from monotherapy up to combinations of five concomitant classes of drugs across each of 10 clinical cohorts defined by age, insulin dependence, and a number of other chronic conditions. A causal deep learning approach developed on such data allows for more personalized evaluation of treatment selection. An average confounder-adjusted reduction in HbA1c of 0.69% [-0.75, -0.65] is observed between patients receiving high vs low ranked treatments across cohorts for which the difference was significant. This method can be extended to explore treatment optimization for other chronic conditions.

Hiding single photons with spread spectrum technology.

We describe a proof-of-principal experiment demonstrating the use of spread spectrum technology at the single photon level. We show how single photons with a prescribed temporal shape, in the presence of interfering noise, may be hidden and recovered.

Nanoscale NMR spectroscopy and imaging of multiple nuclear species.

Nuclear magnetic resonance (NMR) spectroscopy and magnetic resonance imaging (MRI) provide non-invasive information about multiple nuclear species in bulk matter, with wide-ranging applications from basic physics and chemistry to biomedical imaging. However, the spatial resolution of conventional NMR and MRI is limited to several micrometres even at large magnetic fields (>1 T), which is inadequate for many frontier scientific applications such as single-molecule NMR spectroscopy and in vivo MRI of individual biological cells. A promising approach for nanoscale NMR and MRI exploits optical measurements of nitrogen-vacancy (NV) colour centres in diamond, which provide a combination of magnetic field sensitivity and nanoscale spatial resolution unmatched by any existing technology, while operating under ambient conditions in a robust, solid-state system. Recently, single, shallow NV centres were used to demonstrate NMR of nanoscale ensembles of proton spins, consisting of a statistical polarization equivalent to ∼100-1,000 spins in uniform samples covering the surface of a bulk diamond chip. Here, we realize nanoscale NMR spectroscopy and MRI of multiple nuclear species ((1)H, (19)F, (31)P) in non-uniform (spatially structured) samples under ambient conditions and at moderate magnetic fields (∼20 mT) using two complementary sensor modalities.

Electro-optic modulation of single photons.

We use the Stokes photon of a biphoton pair to set the time origin for electro-optic modulation of the wave function of the anti-Stokes photon thereby allowing arbitrary phase and amplitude modulation. We demonstrate conditional single-photon wave functions composed of several pulses, or instead, having Gaussian or exponential shapes.

Observation of optical precursors at the biphoton level.

We describe the observation of a sharp leading-edge spike in a biphoton wave packet that is produced using slow light and measured by two-photon correlation. Using the stationary-phase approximation we characterize this spike as a Sommerfeld-Brillouin precursor resulting from the interference of low- and high-frequency spectral components.

Subnatural linewidth biphotons with controllable temporal length.

This Letter describes the generation of biphotons with a temporal length that can be varied over the range of 50-900 ns, with an estimated subnatural linewidth as small as 0.75 MHz. We make use of electromagnetically induced transparency and slow light in a two-dimensional magneto-optical trap with an optical depth as high as 62. We report a sharp leading edge spike that is a Sommerfeld-Brillouin precursor, as observed at the biphoton level.

Generation of narrow-bandwidth paired photons: use of a single driving laser.

We describe a generator of narrow-band paired photons. A single retroreflected Ti:sapphire laser is used to cool, render transparent, and parametrically pump a cloud of (87)Rb atoms. We attain a paired-photon generation rate into opposing fibers of 600 counts/s with an intensity correlation function that has a width of 5 ns, and violates the Cauchy-Schwartz criteria by a factor of 2000.