Dominik Heider

fourSynergy combines multiple 4C-seq tools using gradient-free optimization to improve interaction calling performance across datasets.

The paper uses single cell RNA-sequencing and genome-wide association study data to identify cell types and processes involved in Barrett's esophagus and esophageal adenocarcinoma risk.

gLinDA analyzes microbiome data using swarm learning to detect differential abundance without sharing raw patient data.

Ensemble-GNN enables federated, ensemble-based graph neural networks for disease module discovery using patient-specific PPI networks and omics data.

The paper introduces Autoturbo-DNA, a framework for training autoencoders tailored to the DNA data storage channel, enabling error correction and constraint-based sequence generation.

ODNA uses machine learning to classify organellar DNA sequences in genome assemblies, achieving high accuracy and outperforming existing methods.

The paper presents a workflow for unsupervised encoding selection and ensemble creation for peptide classification, demonstrating that Pareto frontier pruning and Decision Tree base models perform well across datasets.

This paper demonstrates that transfer learning improves analysis of small single-cell datasets by overcoming data scarcity, batch effects, and technical heterogeneity using large-scale bulk data.

MOSGA 2 provides advanced genomic analysis tools for annotation, validation, and comparative genomics across multiple datasets.

Natrix is a Snakemake-based workflow for preprocessing and analyzing amplicon sequencing data, providing reproducible and scalable processing from quality assessment to taxonomic assignment.

sPLINK enables federated GWAS on distributed datasets without data sharing, preserving privacy and accuracy while being robust to heterogeneity.

The paper introduces γBOriS, a motif-based machine learning tool for identifying oriC sequences in Gammaproteobacteria, and presents BOriS DB, a database containing 25,827 oriC sequences from 1,217 species.

The paper presents four high-quality draft genome assemblies of Cafeteria roenbergensis, enabling evolutionary analysis of heterotrophic protists.

This study compares machine learning methods for protein sequence classification using FCGR encoding, finding that deep neural networks outperform others.