Yuxuan Du

A sparse mixture-of-experts DNA model conditioned on hyperbolic taxonomy embeddings learns better microbial representations than its taxonomy-free twin and comes close to Evo2-7B in probing and fine-tuning at lower compute.

MurineCyto-Det provides a high-resolution, annotated murine BALF cytology dataset for cell segmentation and detection, highlighting challenges in automated analysis.

This paper benchmarks 20 computational methods for multi-omics biomarker discovery in cancer using real and simulated datasets, finding that DeePathNet and Deep-KEGG perform best.

The paper benchmarks alignment strategies for Hi-C reads in metagenomic data, finding BWA MEM -5SP performs best in inter-contig read pairs and binning quality.

METAHICT improves metagenomic Hi-C analysis by integrating binning, reassembly, and downstream tools to recover high-quality MAGs and reduce contamination.

HiCBin improves metagenomic binning using Hi-C contact maps, achieving better accuracy than existing methods in recovering high-quality MAGs.

HiCzin normalizes metagenomic Hi-C data and detects spurious contacts using zero-inflated negative binomial regression, improving bias correction and contig clustering.