Sai Zhang

The paper maps 3D genome organization and DNA methylation across 86,689 single nuclei from 16 human tissues, revealing cell-type-specific variations and discrepancies between methylation and chromatin structure.

AART improves cross-platform proteomic translation by integrating ridge regression and residual learning, enhancing reproducibility and accuracy across multiple diseases and platforms.

ProMeta improves disease diagnosis and prediction from plasma proteomics using a meta-learning framework that overcomes data scarcity with biological priors and interpretable features.

PRSNet-2 predicts complex phenotypes from genotypic data using a hierarchical graph neural network, outperforming baselines and reducing overfitting.

MAPA uses large language models to reduce redundancy in pathway analysis and improve biological interpretation of multi-omics data.

STARNet infers gene regulatory networks from spatial multi-omics data, revealing domain-specific regulatory interactions in tissues.

This study uses a deep learning framework to analyze whole-genome data from ME/CFS patients, identifying risk genes and linking them to disease symptoms and other complex traits.

Prophet is a transformer-based framework that improves disease diagnosis and prediction using proteomics data from the UK Biobank, achieving high accuracy and identifying new protein-disease associations.

The paper introduces predicTTE, a tool for time-to-event prediction in neurological diseases, demonstrating its utility in genetic discovery and personalized medicine.

This paper introduces scPRS, a deep learning model that deconvolves polygenic risk scores at the single-cell level to uncover cellular and molecular disease mechanisms in complex human diseases.

The paper identifies CCDC146 as a modifier of ALS survival through primary cilia dysfunction and shows ASO targeting of CCDC146 rescues ALS phenotypes in human and mouse models.

PRS-Net improves disease prediction by modeling genetic nonlinearity using deep learning and identifies key genes and interactions influencing Alzheimer's and multiple sclerosis risk.

The paper identifies reduced EXOSC2 expression as protective against SARS-CoV-2 infection and clinical COVID-19 by inhibiting viral replication and upregulating antiviral genes.

The paper introduces deepPseudoMSI, a deep learning framework that converts LC-MS data into pseudo-MS images for improved disease diagnosis in precision medicine.

The paper uses systems analysis to identify a molecular network in hypoplastic left heart syndrome, revealing novel factors and mechanisms through integration of genomic and transcriptomic data.

This study combines common and rare variant analyses with single-cell multiomics to identify cell-type-specific genetic risk factors for severe COVID-19, revealing molecular mechanisms and potential therapeutic targets.

This study integrates longitudinal personal exposome and multi-omics data to show dynamic interactions between environmental exposures and internal biomolecules, highlighting key agrochemicals and fungi.

The paper investigates how mitochondrial function affects ALS severity but not risk, identifying specific genetic and expression factors linked to disease progression.

This study identifies 690 candidate ALS genes using a hierarchical Bayesian model, highlighting genetic convergence on the distal axon and functional significance in motor neurons.

This study introduces PULSE, a deep learning model that predicts pseudouridylation sites and characterizes their sequence features using convolutional neural networks.

DeepHINT predicts HIV-1 integration sites and explains their selection using an attention-based deep learning framework, outperforming conventional methods.

GEM reconstructs 3D chromatin structures from Hi-C data using manifold learning and conformational energy, outperforming existing methods and recovering long-range interactions.

DeBooster improves RBP target prediction from CLIP-seq data by reducing false positives and negatives, and enables new biological insights into RBP functions.

TITER predicts translation initiation sites using deep learning on QTI-seq data, outperforming existing methods and identifying sequence signatures for different TIS codons.

ROSE predicts ribosome stalling from ribosome profiling data, showing better performance than conventional models and revealing regulatory factors and functional interactions.