Kenta Nakai

LingoCell interprets gene perturbations using natural language and tissue niches, revealing synergistic disease responses and shared fibroblast programs in cancers.

VINTER is a generative vision-language model that annotates spatial tissues by integrating histology, gene expression, and spatial neighbourhood, outperforming existing methods and transferring across datasets.

SpliceSelectNet predicts splice sites from DNA sequences using a hierarchical Transformer model, achieving state-of-the-art performance and capturing long-range regulatory effects.

This study generates the first spatial transcriptomic atlas of the adult Ciona neural complex, revealing molecular zonation and functional roles of key regions like the neural gland.

DiffusionST improves spatial transcriptomics data quality and identifies spatial domains through a diffusion model and graph convolutional network.

Gene2role embeds genes using multi-hop topological information from signed GRNs to identify topological changes across cell types and quantify module stability.

SCOIGET infers spatial copy number variation distributions using graph neural networks, improving accuracy and efficiency in tumor evolution analysis.

STAIG integrates spatial transcriptomics data with histological images using graph-contrastive learning to analyze spatial domains without prealignment.

The paper compares single-cell transcriptomes across four vertebrate species to identify conserved regulatory patterns and differentiation drivers in retinal cells.

The paper develops a computational method to model interactions between proximal and distal regulators in B cell differentiation, finding that distal regulators fine-tune transcription and modular regulators are involved in chromatin reorganization.

HyGAnno transfers cell type labels from scRNA-seq to scATAC-seq using a graph neural network, improving annotation accuracy and robustness.

DenoiseST is a deep learning method for spatial transcriptomics that improves data imputation, clustering, and identification of functionally variable genes by reducing dropout noise.

This study identifies hysteresis genes in macrophage polarization and shows that AP-1 and CTCF regulate chromatin state changes during de/repolarization.

The paper identifies 11,000 housekeeping CREs (HK-CREs) and shows their reduced activity in cancer, correlating with lower gene expression and poorer patient survival.

DeepBIO is a deep-learning platform that automates biological sequence analysis, offering 42 algorithms and visualization tools for functional prediction and annotation.

The paper generates transgenic mice with prairie vole Oxtr locus to study how long-range genomic regulation affects brain OXT receptor expression and social behaviors.

The paper applies extreme value theory to analyze mutation frequency distributions in cancer genomes, finding a Frechet distribution rather than the Gumbel hypothesis.

The paper develops a computational method to detect microbial contaminants in NGS data and assesses their functional impact on host molecular pathways.

The paper examines the temporal dynamics of histone modifications in mouse dendritic cells after LPS stimulation, revealing distinct patterns and links to transcription factor activity.