Tatsuhiko Tsunoda

QBayMic improves clustering and feature selection in low-signal microbiome data by using quantum-coupled variational Bayes, outperforming classical methods in stability and accuracy.

DBayesCM integrates multi-omics microbiome data using a deep Bayesian model to enable interpretable clustering and relationship prediction across diseases.

vDeepInsight introduces a 3D voxel carrier for tabular data that improves feature neighborhood representation, leading to better performance in gene expression tasks.

This paper improves pan-cancer drug response prediction by addressing class imbalance and shortcut learning in domain adaptation models, achieving better performance in external patient cohorts.

This paper develops a dynamic multimodal model for survival prediction in multiple myeloma using gene expression, lab data, and treatment history, achieving higher accuracy than existing methods.

scHDeepInsight improves immune cell classification in scRNA-seq by using a hierarchical deep learning framework with adaptive loss and image-based features, achieving 93.2% accuracy across seven datasets.

VBayesMM improves identification of microbial-metabolite relationships in high-dimensional microbiome data by using a Bayesian neural network with spike-and-slab priors.

This paper introduces PepCNN, a deep learning model for predicting protein-peptide binding that improves specificity, precision, and AUC over existing methods.

The paper introduces MRep-DeepInsight, a method that improves tabular data analysis by using multiple representations, showing better accuracy than traditional methods on various datasets.

scDeepInsight improves cell-type annotation in scRNA-seq by using supervised deep learning, achieving higher accuracy than existing methods.

This study identifies biallelic PNPLA8 variants causing microcephaly by reducing basal radial glia in cerebral organoids and disrupting phospholipid metabolism.

DRPbind predicts DNA, RNA, and protein-binding residues in disordered protein sequences using SVM and features from evolutionary, physicochemical, and structural data.

DeepInsight-3D predicts patient-specific anticancer drug response using multi-omics data and convolutional neural networks, outperforming existing methods.

The paper analyzes splicing QTLs focusing on coding sequences to identify disease-causal variants associated with GWAS traits through novel integration and selection approaches.

The paper reports in silico structural analysis of ethnic-specific PAX4 variants associated with type 2 diabetes in a Japanese population.

DeepInsight-FS uses a convolutional neural network to select clinically relevant gene subsets from non-image data by converting samples into image-form.

This study uses GWAS to identify Mendelian and non-Mendelian variants associated with retinitis pigmentosa, uncovering new genetic causes and therapeutic targets.

This paper develops a machine learning method to predict drug toxicity by integrating drug targets, functional impact scores, and biological networks, showing it can identify toxic drugs and support safe combination design.

The paper uses empirical Bayes estimation of semi-parametric hierarchical mixture models to analyze GWAS summary statistics and characterize the polygenic architectures of schizophrenia and rheumatoid arthritis.

The paper identifies and characterizes coding and non-coding driver mutations across 2,500 cancer genomes, revealing new candidates and questioning previous findings.