HyFormer: Hybrid Transformer and CNN for Pixel-Level Multispectral Image Land Cover Classification.

To effectively solve the problems that most convolutional neural networks cannot be applied to the pixelwise input in remote sensing (RS) classification and cannot adequately represent the spectral sequence information, we propose a new multispectral RS image classification framework called HyFormer based on Transformer. First, a network framework combining a fully connected layer (FC) and convolutional neural network (CNN) is designed, and the 1D pixelwise spectral sequences obtained from the fully connected layers are reshaped into a 3D spectral feature matrix for the input of CNN, which enhances the dimensionality of the features through FC as well as increasing the feature expressiveness, and can solve the problem that 2D CNN cannot achieve pixel-level classification. Secondly, the features of the three levels of CNN are extracted and combined with the linearly transformed spectral information to enhance the information expression capability, and also used as the input of the transformer encoder to improve the features of CNN using the powerful global modelling capability of the Transformer, and finally the skip connection of the adjacent encoders to enhance the fusion between different levels of information. The pixel classification results are obtained by MLP Head. In this paper, we mainly focus on the feature distribution in the eastern part of Changxing County and the central part of Nanxun District, Zhejiang Province, and conduct experiments based on Sentinel-2 multispectral RS images. The experimental results show that the overall accuracy of HyFormer for the study area classification in Changxing County is 95.37% and that of Transformer (ViT) is 94.15%. The experimental results show that the overall accuracy of HyFormer for the study area classification in Nanxun District is 95.4% and that of Transformer (ViT) is 94.69%, and the performance of HyFormer on the Sentinel-2 dataset is better than that of the Transformer.

Multiplexed PCR-Free Detection of MicroRNAs in Single Cancer Cells Using a DNA-Barcoded Microtrough Array Chip.

MicroRNAs are a class of small RNA molecules that regulate the expression of mRNAs in a wide range of biological processes and are implicated in human health and disease such as cancers. How to measure microRNA profiles in single cells with high throughput is essential to the development of cell-based assays for interrogating microRNA-mediated intratumor heterogeneity and the design of new lab tests for diagnosis and monitoring of cancers. Here, we report on an in situ hybridization barcoding workflow implemented in a sub-nanoliter microtrough array chip for high-throughput and multiplexed microRNA detection at the single cell level. The microtroughs are used to encapsulate single cells that are fixed, permeabilized, and pre-incubated with microRNA detection probes, each of which consists of a capture strand complementary to specific microRNA and a unique reporter strand that can be photocleaved in the microtroughs and subsequently detected by an array of DNA barcodes patterned on the bottom of the microtroughs. In this way, the measurement of reporter strands released from single cells is a surrogate for detecting single-cell microRNA profiles. This approach permits direct measurement of microRNAs without PCR amplification owing to the small volume (<1 nL) of microtroughs. It offers high throughput and high multiplexing capability for evaluating microRNA heterogeneity in single cells, representing a new approach toward microRNA-based diagnosis and monitoring of complex human diseases.

Single-cell microRNA-mRNA co-sequencing reveals non-genetic heterogeneity and mechanisms of microRNA regulation.

Measuring multiple omics profiles from the same single cell opens up the opportunity to decode molecular regulation that underlies intercellular heterogeneity in development and disease. Here, we present co-sequencing of microRNAs and mRNAs in the same single cell using a half-cell genomics approach. This method demonstrates good robustness (~95% success rate) and reproducibility (R2 = 0.93 for both microRNAs and mRNAs), yielding paired half-cell microRNA and mRNA profiles, which we can independently validate. By linking the level of microRNAs to the expression of predicted target mRNAs across 19 single cells that are phenotypically identical, we observe that the predicted targets are significantly anti-correlated with the variation of abundantly expressed microRNAs. This suggests that microRNA expression variability alone may lead to non-genetic cell-to-cell heterogeneity. Genome-scale analysis of paired microRNA-mRNA co-profiles further allows us to derive and validate regulatory relationships of cellular pathways controlling microRNA expression and intercellular variability.

Capture, amplification, and global profiling of microRNAs from low quantities of whole cell lysate.

MicroRNAs (miRNAs) are small non-coding RNAs that control gene expression at the post-transcriptional level via a complex regulatory network that requires genome-wide miRNA profiling to dissect. The patterns of miRNA expression at the genome scale are rich in diagnostic and prognostic information for human diseases such as cancers. This analysis, however, requires multi-step purification of RNAs from large quantities of cells, which is not only time consuming and costly but also challenging in situations where cell numbers are limited. In this study, we report direct capture, amplification, and library preparation of miRNAs from whole cell lysate without the need of pre-purification. As a result, it enables genome-wide miRNA profiling reproducibly with low quantity of cell samples (∼500 hematopoietic cells). Specifically, we conducted a systematic investigation of two key steps - cell lysis for miRNA release and 3' adaptor ligation required for direct miRNA capture and amplification. The obtained expression profile not only distinguishes cell types but also detects individual miRNA alterations in closely related isogenic cell lines. This approach, which is substantially simple as compared to the standard methods because of elimination of the need for RNA purification, is advantageous for the measurement of low quantity samples.

miR-125b promotes MLL-AF9-driven murine acute myeloid leukemia involving a VEGFA-mediated non-cell-intrinsic mechanism.

The hematopoietic stem cell-enriched miR-125 family microRNAs (miRNAs) are critical regulators of hematopoiesis. Overexpression of miR-125a or miR-125b is frequent in human acute myeloid leukemia (AML), and the overexpression of these miRNAs in mice leads to expansion of hematopoietic stem cells accompanied by perturbed hematopoiesis with mostly myeloproliferative phenotypes. However, whether and how miR-125 family miRNAs cooperate with known AML oncogenes in vivo, and how the resultant leukemia is dependent on miR-125 overexpression, are not well understood. We modeled the frequent co-occurrence of miR-125b overexpression and MLL translocations by examining functional cooperation between miR-125b and MLL-AF9 By generating a knock-in mouse model in which miR-125b overexpression is controlled by doxycycline induction, we demonstrated that miR-125b significantly enhances MLL-AF9-driven AML in vivo, and the resultant leukemia is partially dependent on continued overexpression of miR-125b Surprisingly, miR-125b promotes AML cell expansion and suppresses apoptosis involving a non-cell-intrinsic mechanism. MiR-125b expression enhances VEGFA expression and production from leukemia cells, in part by suppressing TET2 Recombinant VEGFA recapitulates the leukemia-promoting effects of miR-125b, whereas knockdown of VEGFA or inhibition of VEGF receptor 2 abolishes the effects of miR-125b In addition, significant correlation between miR-125b and VEGFA expression is observed in human AMLs. Our data reveal cooperative and dependent relationships between miR-125b and the MLL oncogene in AML leukemogenesis, and demonstrate a miR-125b-TET2-VEGFA pathway in mediating non-cell-intrinsic leukemia-promoting effects by an oncogenic miRNA.

microRNA Expression Profiling: Technologies, Insights, and Prospects.

Since the early days of microRNA (miRNA) research, miRNA expression profiling technologies have provided important tools toward both better understanding of the biological functions of miRNAs and using miRNA expression as potential diagnostics. Multiple technologies, such as microarrays, next-generation sequencing, bead-based detection system, single-molecule measurements, and quantitative RT-PCR, have enabled accurate quantification of miRNAs and the subsequent derivation of key insights into diverse biological processes. As a class of ~22 nt long small noncoding RNAs, miRNAs present unique challenges in expression profiling that require careful experimental design and data analyses. We will particularly discuss how normalization and the presence of miRNA isoforms can impact data interpretation. We will present one example in which the consideration in data normalization has provided insights that helped to establish the global miRNA expression as a tumor suppressor. Finally, we discuss two future prospects of using miRNA profiling technologies to understand single cell variability and derive new rules for the functions of miRNA isoforms.