Akira Funahashi

The paper uses live-imaging to study proliferation dynamics of post-WGD HCT116 cell lineages, finding that multipolar chromosome segregation limits proliferative capacity.

This paper introduces a deep learning model to predict gene regulation relationships for both known and unknown transcription factors, achieving higher accuracy for unknown TFs than existing methods.

This paper introduces an unsupervised domain adaptation method for cell segmentation that avoids predefined thresholds, achieving accuracy comparable to supervised models without labeled data.

This paper introduces MaLoN, an unsupervised method for inferring gene regulatory networks from in vivo data, showing improved accuracy over existing methods in yeast and E. coli.

The paper simulates cytokinesis by modeling cytoplasmic flow and cortical elasticity interactions, showing spindle-driven flow creates a pressure minimum that initiates cleavage furrow formation.

This paper develops a deep learning model to automatically predict mouse seminiferous tubule stages from bright-field microscopy images, achieving 98.33% accuracy with a ±1 stage tolerance.

This paper develops a new image-to-image translation model for label-free multiplex microscopic imaging, achieving high performance at both pixel and image levels.

The paper develops FL2-Net, a deep learning method for nuclear segmentation in bright-field microscopy images of mouse embryos, achieving higher birth prediction accuracy than existing methods and experts.

The paper presents a computational framework for analyzing SARS-CoV-2-host interactions to identify drug targets and model disease mechanisms.

The paper uses genome editing to show that calcium signaling pathways, specifically calmodulin-calcineurin-NFAT and calmodulin-calmodulin kinase-CREB, play distinct roles in anhydrobiosis in Pv11 cells.

The paper describes the development of a computational knowledge repository, the COVID-19 Disease Map, which curates SARS-CoV-2 virus-host interaction mechanisms and provides tools for their analysis and application.

The paper develops a deep learning algorithm to predict mouse embryo live birth potential from nuclear structural features in fluorescence images, achieving 83.87% accuracy.

This paper introduces a deep learning-based instance segmentation algorithm, QCA Net, which outperforms existing methods in segmenting nuclei in mouse embryos and provides quantitative developmental metrics.

This study uses convolutional neural networks to predict future cell movement direction from current cell shape images with high accuracy.

This study uses CAGE-seq to analyze transcriptome changes in Pv11 cells during desiccation and rehydration, identifying DEGs and suggesting DNA repair as a recovery mechanism.