Georg Hahn

The paper introduces ROMY, a framework for analyzing multi-omics data with robust association testing, variance-covariance effects, and interaction testing using double machine learning.

This paper quantifies individual dataset contributions to prediction accuracy using cooperative learning in the CAMP study.

This paper develops polygenic hazard score models using lasso for Cox's proportional hazards to predict Alzheimer's free survival, showing improved accuracy and personalized survival curves.

This paper introduces fast algorithms for computing principal components of genomic similarity matrices, including exact methods for specific matrices and an approximate Jaccard matrix with theoretical guarantees.

The paper uses unsupervised cluster analysis to show that omicron variant SARS-CoV-2 sequences are most similar to early 2020 sequences, suggesting long-term circulation.

The paper introduces a smoothed Lasso method for high-dimensional regression and applies it to compute polygenic risk scores for COPD, showing comparable accuracy to FISTA with similar runtime.

This paper introduces a smoothed version of the Lassosum penalty to improve integrated risk models, showing increased accuracy in predicting Alzheimer's and COPD risks compared to the original Lassosum and LDpred2.

The paper uses unsupervised clustering of SARS-CoV-2 genomes to show geographic progression and identify genetic subgroups.

The paper identifies two SARS-CoV-2 genome loci associated with increased COVID-19 mortality risk using logistic regression on whole-genome sequencing data.

This paper introduces a smoothing framework for L1 penalized regression that provides accurate predictions with theoretical guarantees and comparable runtime to existing methods.

locStra efficiently analyzes regional and global population stratification in WGS studies using sparse matrix algebra and sliding window approaches to improve association testing.

The paper analyzes SARS-CoV-2 genomes from Beijing to determine their genetic origin, finding they cluster with European and South(east) Asian samples from 2020.