Krzysztof Kotlarz

The paper introduces a procedure to analyze gene panel data for complex traits, narrowing down SNPs with potential causal effects on protein structure.

This paper compares three feature selection methods for ultra-high-dimensional genomic data, finding that MD-SRA provides the best balance of classification accuracy and computational efficiency.

The paper uses an autoencoder to identify systematic errors in SNP genotype calling from WGS data by analyzing sequence context and quality metrics.

The paper develops a deep learning classifier for bovine mastitis susceptibility using whole genome sequence data, overcoming the p>>n problem through LASSO and deep learning integration.

This paper compares SNP calling workflow parallelisation using plain Bash and Nextflow, finding that multi-process Nextflow is faster but uses more RAM and disk space.

The paper estimates the functional effects of KEGG pathways on bovine stature using SNP summary statistics from GWAS and mixed linear models.