A Deep Learning Model for Predicting Molecular Subtype of Breast Cancer by Fusing Multiple Sequences of DCE-MRI From Two Institutes.

RATIONALE AND OBJECTIVES: To evaluate the performance of deep learning (DL) in predicting different breast cancer molecular subtypes using DCE-MRI from two institutes. MATERIALS AND METHODS: This retrospective study included 366 breast cancer patients from two institutes, divided into training (n = 292), validation (n = 49) and testing (n = 25) sets. We first transformed the public DCE-MRI appearance to ours to alleviate small-data-size and class-imbalance issues. Second, we developed a multi-branch convolutional-neural-network (MBCNN) to perform molecular subtype prediction. Third, we assessed the MBCNN with different regions of interest (ROIs) and fusion strategies, and compared it to previous DL models. Area under the curve (AUC) and accuracy (ACC) were used to assess different models. Delong-test was used for the comparison of different groups. RESULTS: MBCNN achieved the optimal performance under intermediate fusion and ROI size of 80 pixels with appearance transformation. It outperformed CNN and convolutional long-short-term-memory (CLSTM) in predicting luminal B, HER2-enriched and TN subtypes, but without demonstrating statistical significance except against CNN in TN subtypes, with testing AUCs of 0.8182 vs. [0.7208, 0.7922] (p=0.44, 0.80), 0.8500 vs. [0.7300, 0.8200] (p=0.36, 0.70) and 0.8900 vs. [0.7600, 0.8300] (p=0.03, 0.63), respectively. When predicting luminal A, MBCNN outperformed CNN with AUCs of 0.8571 vs. 0.7619 (p=0.08) without achieving statistical significance, and is comparable to CLSTM. For four-subtype prediction, MBCNN achieved an ACC of 0.64, better than CNN and CLSTM models with ACCs of 0.48 and 0.52, respectively. CONCLUSION: Developed DL model with the feature extraction and fusion of DCE-MRI from two institutes enabled preoperative prediction of breast cancer molecular subtypes with high diagnostic performance.

AI-guided histopathology predicts brain metastasis in lung cancer patients.

Brain metastases can occur in nearly half of patients with early and locally advanced (stage I-III) non-small cell lung cancer (NSCLC). There are no reliable histopathologic or molecular means to identify those who are likely to develop brain metastases. We sought to determine if deep learning (DL) could be applied to routine H&E-stained primary tumor tissue sections from stage I-III NSCLC patients to predict the development of brain metastasis. Diagnostic slides from 158 patients with stage I-III NSCLC followed for at least 5 years for the development of brain metastases (Met+, 65 patients) versus no progression (Met-, 93 patients) were subjected to whole-slide imaging. Three separate iterations were performed by first selecting 118 cases (45 Met+, 73 Met-) to train and validate the DL algorithm, while 40 separate cases (20 Met+, 20 Met-) were used as the test set. The DL algorithm results were compared to a blinded review by four expert pathologists. The DL-based algorithm was able to distinguish the eventual development of brain metastases with an accuracy of 87% (p < 0.0001) compared with an average of 57.3% by the four pathologists and appears to be particularly useful in predicting brain metastases in stage I patients. The DL algorithm appears to focus on a complex set of histologic features. DL-based algorithms using routine H&E-stained slides may identify patients who are likely to develop brain metastases from those who will remain disease free over extended (>5 year) follow-up and may thus be spared systemic therapy. © 2024 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

CO Diffusion Study and Spatial and Temporal Variation Modeling during the Construction Period of the Plateau Railroad Tunnel.

In order to investigate the diffusion law of CO gas in the vicinity of the tunnel boring face of the plateau long tunnel, to improve the efficiency of tunnel smoke exhaust, and to derive the spatial-temporal variation model of CO concentration for predicting the concentration of CO at different times and in different cross sections under specific environments, a CO diffusion model of a tunnel in Yunnan was established by using Ansys Fluent Fluid Simulation Software, and the CO transport characteristics under different conditions were simulated by taking the ventilation time, wind speed, and location of the air ducts as the influencing factors. The results show that the wind flows from the mouth of the wind pipe after the wind speed decreases, the diffusion area increases and arrives at the face of the direction of the rebound in the jet stream of new wind, and the return wind under the joint action of the vortex produced obviously, to reach the wind pipe mouth after the tunnel wind flow field, basically tends to stabilize. When the wind pipe mouth was arranged in the arch waist, 20 m away from the boring face, the inlet wind speed was 9 m/s and the ventilation time was 30 min; the CO concentration in the tunnel was reduced to below the maximum allowable concentration value. Moreover, the concentration of CO in the tunnel at the moment of 15 min of ventilation has a nonlinear positive correlation with the change of distance L from the boring face, while at the cross section of the air outlet of the wind pipe L = 20 m, the ventilation time is from 1 to 30 min and the concentration of CO at the cross section has a nonlinear decreasing trend with the ventilation time, which can be deduced according to the different space-time change models.

Dectin-1 stimulation promotes a distinct inflammatory signature in the setting of HIV-infection and aging.

Dectin-1 is an innate immune receptor that recognizes and binds ß-1, 3/1, 6 glucans on fungi. We evaluated Dectin-1 function in myeloid cells in a cohort of HIV-positive and HIV-negative young and older adults. Stimulation of monocytes with ß-D-glucans induced a pro-inflammatory phenotype in monocytes of HIV-infected individuals that was characterized by increased levels of IL-12, TNF-α, and IL-6, with some age-associated cytokine increases also noted. Dendritic cells showed a striking HIV-associated increase in IFN-α production. These increases in cytokine production paralleled increases in Dectin-1 surface expression in both monocytes and dendritic cells that were noted with both HIV and aging. Differential gene expression analysis showed that HIV-positive older adults had a distinct gene signature compared to other cohorts characterized by a robust TNF-α and coagulation response (increased at baseline), a persistent IFN-α and IFN-γ response, and an activated dendritic cell signature/M1 macrophage signature upon Dectin-1 stimulation. Dectin-1 stimulation induced a strong upregulation of MTORC1 signaling in all cohorts, although increased in the HIV-Older cohort (stimulation and baseline). Overall, our study demonstrates that the HIV Aging population has a distinct immune signature in response to Dectin-1 stimulation. This signature may contribute to the pro-inflammatory environment that is associated with HIV and aging.

Assaying Chemical Long-Term Potentiation in Human iPSC-Derived Neuronal Networks.

Impairment of long-term potentiation (LTP) is a common feature of many preclinical models of neurological disorders. Modeling LTP on human induced pluripotent stem cells (hiPSC) enables the investigation of this critical plasticity process in disease-specific genetic backgrounds. Here, we describe a method to chemically induce LTP across entire networks of hiPSC-derived neurons on multi-electrode arrays (MEAs) and investigate effects on neuronal network activity and associated molecular changes.

Spatially resolved single-cell translatomics at molecular resolution.

The precise control of messenger RNA (mRNA) translation is a crucial step in posttranscriptional gene regulation of cellular physiology. However, it remains a challenge to systematically study mRNA translation at the transcriptomic scale with spatial and single-cell resolution. Here, we report the development of ribosome-bound mRNA mapping (RIBOmap), a highly multiplexed three-dimensional in situ profiling method to detect cellular translatome. RIBOmap profiling of 981 genes in HeLa cells revealed cell cycle-dependent translational control and colocalized translation of functional gene modules. We mapped 5413 genes in mouse brain tissues, yielding spatially resolved single-cell translatomic profiles for 119,173 cells and revealing cell type-specific and brain region-specific translational regulation, including translation remodeling during oligodendrocyte maturation. Our method detected widespread patterns of localized translation in neuronal and glial cells in intact brain tissue networks.

AlphaTracker: a multi-animal tracking and behavioral analysis tool.

Computer vision has emerged as a powerful tool to elevate behavioral research. This protocol describes a computer vision machine learning pipeline called AlphaTracker, which has minimal hardware requirements and produces reliable tracking of multiple unmarked animals, as well as behavioral clustering. AlphaTracker pairs a top-down pose-estimation software combined with unsupervised clustering to facilitate behavioral motif discovery that will accelerate behavioral research. All steps of the protocol are provided as open-source software with graphic user interfaces or implementable with command-line prompts. Users with a graphical processing unit (GPU) can model and analyze animal behaviors of interest in less than a day. AlphaTracker greatly facilitates the analysis of the mechanism of individual/social behavior and group dynamics.

Automatic detection of circulating tumor cells and cancer associated fibroblasts using deep learning.

Circulating tumor cells (CTCs) and cancer-associated fibroblasts (CAFs) from whole blood are emerging as important biomarkers that potentially aid in cancer diagnosis and prognosis. The microfilter technology provides an efficient capture platform for them but is confounded by two challenges. First, uneven microfilter surfaces makes it hard for commercial scanners to obtain images with all cells in-focus. Second, current analysis is labor-intensive with long turnaround time and user-to-user variability. Here we addressed the first challenge through developing a customized imaging system and data pre-processing algorithms. Utilizing cultured cancer and CAF cells captured by microfilters, we showed that images from our custom system are 99.3% in-focus compared to 89.9% from a top-of-the-line commercial scanner. Then we developed a deep-learning-based method to automatically identify tumor cells serving to mimic CTC (mCTC) and CAFs. Our deep learning method achieved precision and recall of 94% (± 0.2%) and 96% (± 0.2%) for mCTC detection, and 93% (± 1.7%) and 84% (± 3.1%) for CAF detection, significantly better than a conventional computer vision method, whose numbers are 92% (± 0.2%) and 78% (± 0.3%) for mCTC and 58% (± 3.9%) and 56% (± 3.5%) for CAF. Our custom imaging system combined with deep learning cell identification method represents an important advance on CTC and CAF analysis.

Multimodal charting of molecular and functional cell states via in situ electro-sequencing.

Paired mapping of single-cell gene expression and electrophysiology is essential to understand gene-to-function relationships in electrogenic tissues. Here, we developed in situ electro-sequencing (electro-seq) that combines flexible bioelectronics with in situ RNA sequencing to stably map millisecond-timescale electrical activity and profile single-cell gene expression from the same cells across intact biological networks, including cardiac and neural patches. When applied to human-induced pluripotent stem-cell-derived cardiomyocyte patches, in situ electro-seq enabled multimodal in situ analysis of cardiomyocyte electrophysiology and gene expression at the cellular level, jointly defining cell states and developmental trajectories. Using machine-learning-based cross-modal analysis, in situ electro-seq identified gene-to-electrophysiology relationships throughout cardiomyocyte development and accurately reconstructed the evolution of gene expression profiles based on long-term stable electrical measurements. In situ electro-seq could be applicable to create spatiotemporal multimodal maps in electrogenic tissues, potentiating the discovery of cell types and gene programs responsible for electrophysiological function and dysfunction.

Spatiotemporally resolved transcriptomics reveals the subcellular RNA kinetic landscape.

Spatiotemporal regulation of the cellular transcriptome is crucial for proper protein expression and cellular function. However, the intricate subcellular dynamics of RNA remain obscured due to the limitations of existing transcriptomics methods. Here, we report TEMPOmap-a method that uncovers subcellular RNA profiles across time and space at the single-cell level. TEMPOmap integrates pulse-chase metabolic labeling with highly multiplexed three-dimensional in situ sequencing to simultaneously profile the age and location of individual RNA molecules. Using TEMPOmap, we constructed the subcellular RNA kinetic landscape in various human cells from transcription and translocation to degradation. Clustering analysis of RNA kinetic parameters across single cells revealed 'kinetic gene clusters' whose expression patterns were shaped by multistep kinetic sculpting. Importantly, these kinetic gene clusters are functionally segregated, suggesting that subcellular RNA kinetics are differentially regulated in a cell-state- and cell-type-dependent manner. Spatiotemporally resolved transcriptomics provides a gateway to uncovering new spatiotemporal gene regulation principles.

Integrative in situ mapping of single-cell transcriptional states and tissue histopathology in a mouse model of Alzheimer's disease.

Complex diseases are characterized by spatiotemporal cellular and molecular changes that may be difficult to comprehensively capture. However, understanding the spatiotemporal dynamics underlying pathology can shed light on disease mechanisms and progression. Here we introduce STARmap PLUS, a method that combines high-resolution spatial transcriptomics with protein detection in the same tissue section. As proof of principle, we analyze brain tissues of a mouse model of Alzheimer's disease at 8 and 13 months of age. Our approach provides a comprehensive cellular map of disease progression. It reveals a core-shell structure where disease-associated microglia (DAM) closely contact amyloid-ß plaques, whereas disease-associated astrocyte-like (DAA-like) cells and oligodendrocyte precursor cells (OPCs) are enriched in the outer shells surrounding the plaque-DAM complex. Hyperphosphorylated tau emerges mainly in excitatory neurons in the CA1 region and correlates with the local enrichment of oligodendrocyte subtypes. The STARmap PLUS method bridges single-cell gene expression profiles with tissue histopathology at subcellular resolution, providing a tool to pinpoint the molecular and cellular changes underlying pathology.

Quantum Controlled Cold Scattering Challenges Theory.

Our previous rotationally inelastic cold scattering experiments between state prepared D2 (v = 2, j = 2, m = 0) and He disagreed with theory, raising serious concerns about either our understanding of the anisotropic potential or the accuracy of the measurement. To further interrogate interactions between molecular hydrogen and atomic helium, we study the Δj = 1and Δj = 2 rotational relaxation of HD (v = 2, j = 2, m = 0) by collision with He. The two rotational transitions probe different anisotropic components of the van der Waals potential. Our state resolved scattering study shows that these two transitions are mediated by two different shape resonances l = 1 for Δj = 1 and l = 2 for Δj = 2. The strong l = 1 resonance dominates the Δj = 1 scattering, agreeing with theory. However, the dominance of the weaker l = 2 resonance in the Δj = 2 transition, which matches our earlier D2-He result, contradicts theoretical calculations. The continued contradiction, when we expect one-to-one correspondence between our stereodynamically controlled scattering experiment and theoretical calculations, makes us question the accuracy of the weaker anisotropic part of the H2-He interaction potential.

Resonant cold scattering of highly vibrationally excited D2 with Ne.

To accurately map weak D2-Ne long-range interactions, we have studied rotationally inelastic cold scattering of D2 prepared in the vibrationally excited (v = 4) and rotationally aligned (j = 2, m) quantum state within the moving frame of a supersonically expanded mixed molecular beam. In contrast to earlier high energy D2-Ne collision experiments, the (j = 2 → j' = 0) cold scattering produced highly symmetric angular distributions that strongly suggest a resonant quasi-bound collision complex that lives long enough to make a few rotations. Our partial wave analysis indicates that the scattering dynamics is dominated by a single resonant l = 2 orbital, even in the presence of a broad temperature (0-5 K) distribution that allows incoming orbitals up to l = 5. The dominance of a single orbital suggests that the resonant complex stabilizes through the coupling of the internal (j = 2) and orbital (l = 2) angular momentum to produce a total angular momentum of J = 0 for the D2-Ne complex.

Study of the Diffusion Law of Harmful Gases in Tunnel Construction on Plateaus and Optimization of Ventilation Parameters.

To explore the diffusion law of harmful gases near the excavation face in tunnel construction on plateaus, select the optimal ventilation parameters, and improve the ventilation efficiency, Ansys Fluent was used to set the environmental parameters according to the highland, where the tunnel is located, and simulate and curve fit the diffusion phenomenon of CO and NO2 based on the fluid control equations and species transport model. The effects of air velocity, ventilation time, and duct position on the diffusion law of harmful gases were analyzed, from which the optimal ventilation conditions were selected, and the ventilation effects under the optimal conditions were compared with those under the original conditions. The study shows that after the fresh airflow passes through the outlet, part of it flows out along the tunnel wall toward the cave entrance, and the other part interacts with the return air to form a vortex; when the air supply speed is 10 m/s and the distance from the duct outlet to the excavation face is 25 m, the maximum concentration of harmful gases decreases by 92-99% after 20 min of ventilation compared with that before optimization and the smoke exhaust efficiency increases by 2.5% per minute.

Development of a platform to investigate long-term potentiation in human iPSC-derived neuronal networks.

Impairment of long-term potentiation (LTP) is a common feature of many pre-clinical models of neurological disorders; however, studies in humans are limited by the inaccessibility of the brain. Human induced pluripotent stem cells (hiPSCs) provide a unique opportunity to study LTP in disease-specific genetic backgrounds. Here we describe a multi-electrode array (MEA)-based assay to investigate chemically induced LTP (cLTP) across entire networks of hiPSC-derived midbrain dopaminergic (DA) and cortical neuronal populations that lasts for days, allowing studies of the late phases of LTP and enabling detection of associated molecular changes. We show that cLTP on midbrain DA neuronal networks is largely independent of the N-methyl-D-aspartate receptor (NMDAR) and partially dependent on brain-derived neurotrophic factor (BDNF). Finally, we describe activity-regulated gene expression induced by cLTP. This cLTP-MEA assay platform will enable phenotype discovery and higher-throughput analyses of synaptic plasticity on hiPSC-derived neurons.

Coherent Preparation of Highly Vibrating and Rotating D2 Molecules.

Highly vibrationally and rotationally excited hydrogen molecules are of immense interest for understanding and modeling the physics and chemistry of the cold interstellar medium. Using a sequence of two Stark-induced adiabatic Raman passages, we demonstrate the preparation of rotationally excited D2 molecules in the fourth excited vibrational level within its ground electronic state. The nearly complete population transfer to the target state is confirmed by observing both the threshold behavior as a function of the laser power and the depletion of the intermediate level. The vibrational excitation reported here opens new possibilities in the study of the much debated four-center reaction between a pair of hydrogen molecules. Additionally, these rovibrationally excited molecules could be potentially used to generate the high-intensity D- ion beams considered essential for D-T thermonuclear fusion by enhancing the cross section for dissociative electron attachment by 5 orders of magnitude compared to that of the ground state.

Anisotropic dynamics of resonant scattering between a pair of cold aligned diatoms.

The collision dynamics between a pair of aligned molecules in the presence of a partial-wave resonance provide the most sensitive probe of the long-range anisotropic forces important to chemical reactions. Here we control the collision temperature and geometry to probe the dynamics of cold (1-3 K) rotationally inelastic scattering of a pair of optically state-prepared D2 molecules. The collision temperature is manipulated by combining the gating action of laser state preparation and detection with the velocity dispersion of the molecular beam. When the bond axes of both molecules are aligned parallel to the collision velocity, the scattering rate drops by a factor of 3.5 as collision energies >2.1 K are removed, suggesting a geometry-dependent resonance. Partial-wave analysis of the measured angular distribution supports a shape resonance within the centrifugal barrier of the l = 2 incoming orbital. Our experiment illustrates the strong anisotropy of the quadrupole-quadrupole interaction that controls the dynamics of resonant scattering.

Non-recursive transport of intensity phase retrieval with the transport of phase.

The transport of intensity equation (TIE) is a non-interferometric phase retrieval method that originates from the imaginary part of the Helmholtz equation and is equivalent to the law of conservation of energy. From the real part of the Helmholtz equation, the transport of phase equation (TPE), which represents the Eikonal equation in the presence of diffraction, can be derived. The amplitude and phase for an arbitrary optical field should satisfy these coupled equations simultaneously during propagation. In this work, the coupling between the TIE and TPE is exploited to improve the phase retrieval solutions from the TIE. Specifically, a non-recursive fast Fourier transform (FFT)-based phase retrieval method using both the TIE and TPE is demonstrated. Based on the FFT-based TIE solution, a correction factor calculated by the TPE is introduced to improve the phase retrieval results.

Use of structured light in 3D reconstruction of transparent objects.

A simple non-interferometric incoherent light ray propagation model is introduced to perform three-dimensional profiling of transparent objects with typical thicknesses of the order of mm to cm by analyzing the distorted captured image behind the object. A two-dimensional cosine fringe is used as the incident reference image, whose periodicity is markedly altered by the shape of the object. By monitoring the local change in the period, the surface profile is simulated and optimized to achieve minimal error with experimental data and thus determine the final morphology. Our proposed method is simple, robust, straightforward, and single-shot, and can be used with coherent or incoherent illumination. Its feasibility for more complex applications is verified experimentally through rigorous error calculation. Moreover, the application of this technique for arbitrary transparent objects is theoretically attainable and promising.