Jianhua Xuan

This paper proposes MAE-UNETR++ for 3-D lung nodule segmentation, showing that MAE pretraining outperforms random initialization and Decathlon transfer learning in low-data scenarios.

ChIP-BIT2 detects weak transcription factor binding events by integrating ChIP-seq sample and input data using a Bayesian approach.

The paper develops ChIP-GSM, a Bayesian method to infer cis-regulatory modules from ChIP-seq data, showing improved performance in predicting regulatory element activity.

BICORN integrates transcription factor binding and gene expression data to infer sparse cis-regulatory modules and their regulatory networks.

MSIGNET identifies significant gene subnetworks using Metropolis sampling, showing better performance than existing tools on simulated data and providing biologically meaningful results in Parkinson's and ovarian cancer studies.