Sheng Li

LongVerse uses AI agents to generate haplotype-resolved DNA methylation calls from Nanopore and PacBio reads, achieving high agreement with bisulfite sequencing and reducing errors compared to existing methods.

NANOME improves DNA methylation detection using nanopore long-read sequencing by integrating multiple tools and enhancing precision and accuracy in human cell lines.

The paper introduces TAVAC, a metric to quantify overfitting and interpretation reproducibility in Vision Transformer models using attention consistency across training and validation data.

The paper develops a computational approach to model core gene regulatory networks in IDH mutant AML, identifying key regulators like DNMT1 and E2F1 supported by patient survival data.

The paper shows that changes in homologous chromosome pairing at the Stat92E locus are required for transcriptional downregulation during Drosophila stem cell differentiation.

The paper evaluates DNA methylation calling tools for Oxford Nanopore sequencing using human data and provides an online database for comparing tool performance.

This paper introduces HBA-DEALS, a method that simultaneously identifies differential expression and splicing, showing genes can have differential expression, splicing, both, or neither.

ChIA-PIPE is a fully automated pipeline for analyzing and visualizing ChIA-PET data, including quality assessment and visualization tools.

Casilio-ME enhances CRISPR-based DNA demethylation by coupling TET1 with DNA repair factors to activate methylation-silenced genes.

This study improves the detection of m6A in bacterial DNA using nanopore sequencing with a binary classifier, highlighting the importance of read quality and comparing results to PacBio methylation calls.

This paper develops machine learning models to predict the primary site of origin and molecular subtype of metastatic cancers using gene expression data from TCGA, achieving high accuracy across multiple cancer types.