Mingyao Li

This paper demonstrates subcellular mass spectrometry imaging using standard hardware and open-source software, enabling accessible spatial metabolomics with subcellular resolution.

Pixel2Gene improves spatial gene expression reconstruction and prediction by integrating histology images with spatial transcriptomics data.

REMAP reconstructs spatial tissue architecture from scRNA-seq using deep learning and ST references, revealing microenvironmental features and cancer subtypes.

The paper maps human kidney development using single-cell RNA sequencing and spatial transcriptomics, revealing fate decisions influenced by the local microenvironment.

This study develops a method to sequence individual human DRG neuron somas, identifies 16 neuronal types, and reveals human-specific nociceptor types and potential drug targets.

This paper introduces Patho-DBiT, a method for spatial transcriptomics in FFPE tissues, enabling RNA biology analysis including splicing, variants, and regulatory networks in clinical samples.

DeepCombat removes batch effects from neuroimaging data using a deep learning approach, outperforming existing methods in preserving biological variation.

The paper uses spatial single-cell analysis to uncover molecular and structural development of human fetal cortex, revealing early layer and area specification.

The paper introduces NEST, a framework for testing brain-phenotype associations within functional networks, applied to neurodevelopmental imaging data.

The paper identifies CD200 as a specific marker for vascular smooth muscle cells and their derived cells in atherosclerosis using CITE-seq and validates it across multiple techniques.

The paper uses multimodal single-cell profiling to identify diverse monocyte subtypes and links them to cardiovascular disease risk factors.

This study identifies differentially expressed genes and allele-specific expression in macular tissues across AMD stages, highlighting the RPE/choroid's role in disease pathophysiology.

TESLA annotates tumor microenvironment cells and detects spatial transcriptome differences using spatial transcriptomics data.

SpaDecon improves cell-type deconvolution in spatial transcriptomics using semi-supervised learning, outperforming existing methods on real and pseudo datasets.

Multi-omic and spatial profiling of 81 human kidney samples resolves four tissue microenvironments; the fibrotic one's gene signature is reported to beat histopathology at predicting outcome.

MuSiC2 improves cell type proportion estimation in bulk RNA-seq data from different clinical conditions using an iterative method, outperforming traditional approaches.

Modelling all spots of a section jointly with a Vision Transformer, HisToGene predicts expression from H&E images better than a re-implemented ST-Net and adds 4x super-resolution prediction.

This paper introduces CarDEC, a deep learning model that simultaneously corrects batch effects, denoises, and clusters scRNA-seq data, outperforming existing methods in multiple evaluations.

LIQA quantifies isoform expression and detects differential alternative splicing in long-read RNA-seq data, outperforming existing methods in low-coverage and biased read scenarios.

This paper introduces BSCET, a method that detects cell-type-specific allelic expression imbalance using bulk RNA-seq data integrated with scRNA-seq-derived cell-type information, identifying genes related to type 2 diabetes progression.

This paper introduces SpaGCN, a graph convolutional network that identifies spatial domains and spatially variable genes by integrating gene expression, spatial location, and histology in spatial transcriptomics data.

ItClust improves clustering and cell type classification in scRNA-seq by using iterative transfer learning with neural networks, outperforming existing methods across diverse datasets.

This paper introduces DESC, a deep learning method for single-cell clustering that improves accuracy and removes batch effects.

The paper uses genome-edited human kidney organoids and single-cell transcriptomics to study APOL1 risk variants' effects on kidney disease mechanisms.

LinkedSV detects mosaic structural variants from linked-read sequencing data, outperforming existing tools and identifying SVs missed by conventional sequencing methods.

ASEP detects gene-level allele-specific expression in populations using RNA-seq, identifying differential ASE between conditions and extending findings to cardiometabolic traits.

MuSiC deconvolves bulk RNA-seq data using single-cell RNA-seq reference to estimate cell type composition, outperforming existing methods in tissues with closely related cell types.

The paper introduces DESCEND, a method to deconvolve true gene expression distributions from noisy scRNA-seq data, improving estimates of dispersion and burstiness while adjusting for covariates.

SAVER improves gene expression estimates in scRNA-seq by imputing zeros and enhancing low/moderate expression genes, improving downstream analysis accuracy.