Robust probabilistic modeling for single-cell multimodal mosaic integration and imputation via scVAEIT.

Recent advances in single-cell technologies enable joint profiling of multiple omics. These profiles can reveal the complex interplay of different regulatory layers in single cells; still, new challenges arise when integrating datasets with some features shared across experiments and others exclusive to a single source; combining information across these sources is called mosaic integration. The difficulties lie in imputing missing molecular layers to build a self-consistent atlas, finding a common latent space, and transferring learning to new data sources robustly. Existing mosaic integration approaches based on matrix factorization cannot efficiently adapt to nonlinear embeddings for the latent cell space and are not designed for accurate imputation of missing molecular layers. By contrast, we propose a probabilistic variational autoencoder model, scVAEIT, to integrate and impute multimodal datasets with mosaic measurements. A key advance is the use of a missing mask for learning the conditional distribution of unobserved modalities and features, which makes scVAEIT flexible to combine different panels of measurements from multimodal datasets accurately and in an end-to-end manner. Imputing the masked features serves as a supervised learning procedure while preventing overfitting by regularization. Focusing on gene expression, protein abundance, and chromatin accessibility, we validate that scVAEIT robustly imputes the missing modalities and features of cells biologically different from the training data. scVAEIT also adjusts for batch effects while maintaining the biological variation, which provides better latent representations for the integrated datasets. We demonstrate that scVAEIT significantly improves integration and imputation across unseen cell types, different technologies, and different tissues.

Fermi-Level Depinning in Metal/Ge Junctions by Inserting a Carbon Nanotube Layer.

Germanium (Ge)-based devices are recognized as one of the most promising next-generation technologies for extending Moore's law. However, one of the critical issues is Fermi-level pinning (FLP) at the metal/n-Ge interface, and the resulting large contact resistance seriously degrades their performance. The insertion of a thin layer is one main technique for FLP modulation; however, the contact resistance is still limited by the remaining barrier height and the resistance induced by the insertion layer. In addition, the proposed depinning mechanisms are also controversial. Here, the authors report a wafer-scale carbon nanotube (CNT) insertion method to alleviate FLP. The inserted conductive film reduces the effective Schottky barrier height without inducing a large resistance, leading to ohmic contact and the smallest contact resistance between a metal and a lightly doped n-Ge. These devices also indicate that the metal-induced gap states mechanism is responsible for the pinning. Based on the proposed technology, a wafer-scale planar diode array is fabricated at room temperature without using the traditional ion-implantation and annealing technology, achieving an on-to-off current ratio of 4.59 × 104 . This work provides a new way of FLP modulation that helps to improve device performance with new materials.

Improving protected area effectiveness through consideration of different human-pressure baselines.

Previous assessments of the effectiveness of protected areas (PAs) focused primarily on changes in human pressure over time and did not consider the different human-pressure baselines of PAs, thereby potentially over- or underestimating PA effectiveness. We developed a framework that considers both human-pressure baseline and change in human pressure over time and assessed the effectiveness of 338 PAs in China from 2010 to 2020. The initial state of human pressure on PAs was taken as the baseline, and changes in human pressure index (HPI) were further analyzed under different baselines. We used the random forest models to identify the management measures that most improved effectiveness in resisting human pressure for the PAs with different baselines. Finally, the relationships between the changes in the HPI and the changes in natural ecosystems in PAs were analyzed with different baselines. Of PAs with low HPI baselines, medium HPI baselines, and high HPI baselines, 76.92% (n=150), 11.11% (n=12), and 22.86% (n=8) , respectively, showed positive effects in resisting human pressure. Overall, ignoring human-pressure baselines somewhat underestimated the positive effects of PAs, especially for those with low initial human pressure. For PAs with different initial human pressures, different management measures should be taken to improve effectiveness and reduce threats to natural ecosystems. We believe our framework is useful for assessing the effectiveness of PAs globally, and we recommend it be included in the Convention on Biological Diversity Post-2020 Strategy.

Las evaluaciones previas de la efectividad de las áreas protegidas (AP) se han enfocado principalmente en los cambios de las presiones humanas con el tiempo y no han considerado las diferentes líneas base de las presiones humanas en las AP, por lo que potencialmente han sobrestimado o subestimado su efectividad. Desarrollamos un marco de trabajo que considera las líneas base de presión humana y los cambios de las presiones humanas con el tiempo y evaluamos a la efectividad de 338 AP en China entre 2010 y 2020. Consideramos el estado inicial de la presión humana en las AP como la línea base y analizamos los cambios en el índice de presión humana (IPH) bajo diferentes líneas base. Utilizamos modelos de bosque aleatorio para identificar las medidas de gestión que más aumentaron la efectividad de la resistencia a las presiones humanas en las AP con líneas base diferentes. Finalmente, analizamos con diferentes líneas base las relaciones entre los cambios en el IPH y los cambios en los ecosistemas naturales de las AP. De las AP con líneas base de IPH bajas, medianas y altas, 76.92% (n=150), 11.11% (n=12) y 22.86% (n=8), respectivamente, mostraron efectos positivos de resistencia a las presiones humanas. En general, si ignoramos las líneas base de las presiones humanas, se subestiman los efectos positivos de las AP de una u otra manera, especialmente aquellas con poca presión humana al inicio. En el caso de las AP que al inicio tienen diferentes presiones humanas, se deben tomar diferentes medidas de gestión para mejorar la efectividad y reducir las amenazas a los ecosistemas naturales. Creemos que nuestro marco de trabajo sirve para evaluar la efectividad mundial de las AP y recomendamos que se incluya en la Estrategia Post-2020 de la Convención sobre la Diversidad Biológica. Mejoría de la Efectividad de un Área Protegida al Considerar Diferentes Líneas Base de Presión Humana.

Simultaneous inference for generalized linear models with unmeasured confounders.

Tens of thousands of simultaneous hypothesis tests are routinely performed in genomic studies to identify differentially expressed genes. However, due to unmeasured confounders, many standard statistical approaches may be substantially biased. This paper investigates the large-scale hypothesis testing problem for multivariate generalized linear models in the presence of confounding effects. Under arbitrary confounding mechanisms, we propose a unified statistical estimation and inference framework that harnesses orthogonal structures and integrates linear projections into three key stages. It begins by disentangling marginal and uncorrelated confounding effects to recover the latent coefficients. Subsequently, latent factors and primary effects are jointly estimated through lasso-type optimization. Finally, we incorporate projected and weighted bias-correction steps for hypothesis testing. Theoretically, we establish the identification conditions of various effects and non-asymptotic error bounds. We show effective Type-I error control of asymptotic $z$-tests as sample and response sizes approach infinity. Numerical experiments demonstrate that the proposed method controls the false discovery rate by the Benjamini-Hochberg procedure and is more powerful than alternative methods. By comparing single-cell RNA-seq counts from two groups of samples, we demonstrate the suitability of adjusting confounding effects when significant covariates are absent from the model.

Decoupling convolution network for characterizing the metastatic lymph nodes of breast cancer patients.

PURPOSE: The dual-energy computed tomography (DECT) technique is an emerging imaging tool that can better characterize material features and has the potential to be a noninvasive means of predicting lymph node metastasis. The purpose of this study was to establish a DECT-specified quantitative approach based on a neural network to characterize the sentinel lymph node (SLN). METHODS: With IRB approval, we retrospectively collected a total of 229 patients (100/229 metastasis) with biopsy proven breast cancer in this study. The chest and axillary spectral CT examinations were performed prior to the axillary lymph node (ALN) surgery. A decoupling convolution network with 11 ROIs from sequential keV (40 to 140 keV with 10 keV increment) was proposed to explicitly extract the spectral and spatial features in a DECT to predict the lymph node status. Focal loss was introduced as the loss function. The metric of the slope of the spectral Hounsfield unit curve measured at the venous phase was used as the baseline approach in comparison to our approach. In additional, a logistic model with radiomic features was also compared to our approach. The area under ROC curve (AUC) was used as the figure of merit to evaluate the classification performance. RESULTS: By introducing spectral convolution and focal loss, AUC on test set could be improved by 0.15 and 0.01 separately. Compared to the slope of the spectral curve with the average AUC of 0.611 and radiomic model with AUC of 0.825, the proposed approach demonstrates a considerably better performance, with test set AUC value of 0.837, by using decoupling spectral and spatial convolution together with focal loss function. CONCLUSIONS: We presented a new decoupling neural network based quantification method for DECT analysis, which might have potential as a noninvasive tool to predict metastasis lymph node status for breast cancer in clinical practice.

Flexible 64 × 64 Pixel AMOLED Displays Driven by Uniform Carbon Nanotube Thin-Film Transistors.

Carbon nanotube (CNT) thin-film transistors are expected to be promising for use in flexible electronics including flexible and transparent integrated circuits and in wearable chemical and physical sensors and for driving the circuits of flexible display panels. However, current devices based on CNT channels suffer from poor performance uniformity and low manufacturing yield; therefore, they are still far from being practical. This is usually caused by nonuniform deposition of the semiconducting CNTs and the rough surface of flexible substrates. Here, we report CNT thin-film transistors (TFTs) driving a flexible 64 × 64 pixel active matrix light-emitting diode display (AMOLED) by improving the formation of uniform CNT films and developing a new pretreatment technique for flexible substrates. The achieved AMOLED has uniform brightness and a high yield of 99.93% in its 4096 pixels. More than 8000 TFTs with high-purity semiconducting CNTs as the channel material show an average on-off current ratio of ∼107 and a carrier mobility of 16 cm2 V-1 s-1. The standard deviations of the on-state current and the carrier mobility are 4.1 and 6.5%, respectively. Our result shows that the panel driven by high-purity semiconducting CNTs is a promising strategy for the development of next-generation flexible, large-area displays.

Ultrahigh-performance transparent conductive films of carbon-welded isolated single-wall carbon nanotubes.

Single-wall carbon nanotubes (SWCNTs) are ideal for fabricating transparent conductive films because of their small diameter, good optical and electrical properties, and excellent flexibility. However, a high intertube Schottky junction resistance, together with the existence of aggregated bundles of SWCNTs, leads to a degraded optoelectronic performance of the films. We report a network of isolated SWCNTs prepared by an injection floating catalyst chemical vapor deposition method, in which crossed SWCNTs are welded together by graphitic carbon. Pristine SWCNT films show a record low sheet resistance of 41 ohm □-1 at 90% transmittance for 550-nm light. After HNO3 treatment, the sheet resistance further decreases to 25 ohm □-1. Organic light-emitting diodes using this SWCNT film as anodes demonstrate a low turn-on voltage of 2.5 V, a high current efficiency of 75 cd A-1, and excellent flexibility. Investigation of isolated SWCNT-based field-effect transistors shows that the carbon-welded joints convert the Schottky contacts between metallic and semiconducting SWCNTs into near-ohmic ones, which significantly improves the conductivity of the transparent SWCNT network. Our work provides a new avenue of assembling individual SWCNTs into macroscopic thin films, which demonstrate great potential for use as transparent electrodes in various flexible electronics.