Lifei Wang

This paper quantifies shared information between histomorphology and spatial transcriptomics using trajectory pseudotime as a mediating variable across multiple organs.

PertCurve models continuous transcriptional response trajectories from CRISPR screens, improving prediction of genetic modulations by capturing diverse perturbation strengths and downstream gene behaviors.

InClust+ extends the inClust framework to handle multimodal data integration, imputation, and cross-modal generation, showing performance comparable to existing tools.

inClust is a clustering framework that integrates multi-source data using variational autoencoder and vector arithmetic, performing well in various clustering tasks.

scCapsNet is a deep learning model for single-cell RNA data analysis that provides interpretable feature extraction and accurate cell type classification.

scCapsNet-mask automates the cell type classification process of scCapsNet by using a mask, maintaining high accuracy and interpretability.

This paper introduces a deep learning model called NGNC to reconstruct gene regulatory networks and infer gene dynamics from time series scRNA-seq data, achieving state-of-the-art results.

The paper introduces MultiCapsNet, a deep learning model that integrates diverse data sources with higher interpretability, achieving comparable performance to existing models with fewer parameters.