Sehyun Oh

MetaHarmonizer: robust biomedical metadata harmonization and a contamination control for inflated LLM performance on public benchmarks

MetaHarmonizer improves metadata harmonization and addresses LLM contamination in biomedical data, outperforming existing methods on benchmark tasks.

Large-scale Manual Curation and Harmonization of Metadata from Metagenomic and Cancer Genomic Repositories: Challenges and Solutions

The paper curates and harmonizes metadata from metagenomic and cancer genomic repositories to improve data reusability and cross-study integration.

Commonly used compositional data analysis implementations are not advantageous in microbial differential abundance analyses benchmarked against biological ground truth

The paper benchmarks differential abundance methods using microbiome datasets with known biological ground truth, finding that compositional methods are less effective than commonly used non-compositional approaches.

Inverting the model of genomics data sharing with the NHGRI Genomic Data Science Analysis, Visualization, and Informatics Lab-space (AnVIL)

The paper describes AnVIL, a cloud-based platform that enables genomic data analysis without data transfer, improving security, scalability, and collaboration.

GenomicSuperSignature: interpretation of RNA-seq experiments through robust, efficient comparison to public databases

GenomicSuperSignature compares new RNA-seq datasets to public archives for rapid interpretation using precomputed axes of variation.

HGNChelper: Identification and correction of invalid gene symbols for human and mouse

HGNChelper identifies and corrects invalid gene symbols in human and mouse data by leveraging authoritative databases and addressing spreadsheet-related errors.

Reliable analysis of clinical tumor-only whole exome sequencing data

The paper presents a workflow for allele-specific CNA analysis and SNV classification from tumor-only WES data, showing high concordance with gold standard methods.