Yutaka Saitō

pLM-bVAE: protein language model-enhanced binary representation improves quantum annealing-based protein sequence design

This paper introduces pLM-bVAE, a binary variational autoencoder enhanced with protein language model representations, showing improved sequence reconstruction and design performance in quantum annealing-based protein sequence design.

Engineering a highly active thermophilic F1-ATPase by homolog-guided exploration and machine-learning-assisted prioritization

This paper engineers a thermophilic F1-ATPase with higher catalytic activity using homolog-guided mutation identification and machine learning, finding a 1.8-fold increase in rotation rate.

Characterization of Porcine Antibodies from Sequence Repertoire and Structural Data

This study characterizes porcine antibodies using sequence and structural data, identifying unique features that may affect CDR-H3 properties and humanization potential.

Evaluating Lightweight and Full Fine-Tuning Strategies Against Classical Machine Learning for Protein Function Prediction

This study compares different fine-tuning strategies and classical machine learning approaches for protein function prediction using benchmark datasets.

Fitness translocation: improving variant effect prediction with biologically-grounded data augmentation

This paper introduces fitness translocation, a data augmentation method for variant effect prediction that improves model performance using homologous protein fitness data.

Exploring diverse routes to high-affinity-antibody variable domains through deep-sequencing-informed machine learning

This study uses deep sequencing and machine learning to optimize antibody discovery by analyzing training data and search space design for high-affinity VHHs.

Data-efficient protein mutational effect prediction with weak supervision by molecular simulation and protein language models

The paper presents a method that improves protein mutational effect prediction with limited experimental data by combining molecular simulation and protein language models.

Evotuning protocols for Transformer-based variant effect prediction on multi-domain proteins

This paper introduces evotuning protocols for Transformer-based variant effect prediction, showing improved performance by incorporating domain architecture information from homologous sequences.

Comprehensive epigenome characterization reveals diverse transcriptional regulation across human vascular endothelial cells

The paper characterizes the transcriptomic and epigenomic landscape of nine human endothelial cell types, identifying EC-specific enhancers and marker genes, and highlights the role of HOX genes in vascular development and disease.