Degui Zhi

This paper introduces MetaRNN and MetaRNN-indel, deep learning models that differentiate rare pathogenic from benign variants, outperforming existing methods and providing interpretable scores.

The paper introduces iGWAS, a method using self-supervised deep learning to extract phenotypes from medical images for GWAS, finding new genetic loci related to retinal vessel development.

The paper introduces mcPBWT, an algorithm for efficiently scanning composite haplotype patterns in PBWT data, identifying crossover events and gene-conversion tracts.

The paper introduces d-PBWT, a dynamic version of the PBWT data structure, enabling efficient insertion and deletion of haplotypes while maintaining PBWT functionality.

The paper infers genealogical history and social behaviors from IBD segments in UK Biobank data, revealing population-scale genetic connections and regional family preferences.

Meta-Align improves DNA sequence alignment by incorporating quality scores and handling frameshift errors using an HMM-based approach, outperforming TBLASTX in simulated benchmarks.

This study develops geneEXPLORER, a model that predicts gene expression using long-range DNA methylation effects, showing it can accurately predict clinical phenotypes in cancer.

HaploHide hides individual haplotypes within population haplotypes to protect privacy during genetic data sharing without changing infrastructure.

This paper introduces Gene2Vec, a gene embedding method using co-expression data from GEO databases, showing improved pathway recovery and gene interaction prediction.

RaPID detects identity by descent in large biobank-scale cohorts using a randomized positional BWT approach, achieving linear scaling and high accuracy.

The paper develops a novel HWE filtering strategy to improve variant calling quality control in large-scale sequencing by optimizing p-value cutoffs based on site frequency spectra.