Jan A. Freudenthal

The paper introduces permGWAS, a permutation-based GWAS method that improves significance thresholds for skewed phenotypes and reduces false discovery rates compared to Bonferroni correction.

This paper proposes local convolutional neural networks for genomic prediction, showing improved performance over traditional methods in most cases, especially with medium complexity traits and large datasets.

The paper benchmarks seven chloroplast genome assembly tools, finding significant performance differences and assembling 20 novel chloroplast genomes from 105 datasets.

GWAS-Flow is a GPU-accelerated framework for efficient permutation-based GWAS, outperforming custom scripts in speed without loss of accuracy.