Lixin Cheng

This paper introduces scCaT, a deep learning framework combining capsulating architecture and Transformer for sepsis diagnosis using single-cell RNA sequencing data and transferring to bulk RNA data, achieving high accuracy and explainability.

The paper evaluates 10 software tools for computing AUPRC and finds they produce conflicting and overly-optimistic results.

This study shows that targeting adipocyte ESRRA promotes osteogenesis and vascular formation in adipocyte-rich bone marrow by modulating signaling pathways and stem cell fate.

DeepLocRNA predicts RNA subcellular localization using transfer learning from pre-trained RBP-binding models, achieving state-of-the-art performance across human and mouse RNA.

The paper introduces COMSV, a method for identifying structural variations in cancer samples, showing high sensitivity and specificity and detecting hundreds of novel SVs per sample.

The paper identifies circular RNA competing endogenous gene pairs (cceGPs) as predictive indicators for serous ovarian cancer using a novel method and validates their predictive capacity in independent cohorts.

The paper develops a CRISPR on-target activity prediction model called GNL using Bayesian Ridge Regression, showing improved cross-dataset and cross-species performance.

The paper identifies and characterizes moonlighting long non-coding RNAs using a statistical algorithm integrating protein interactome and functional data, revealing distinct sequence and localization features.