Bioinformatics and Omics Data Science

High-throughput technologies and multimodal health data produce vast, interconnected datasets vital for decoding human pathology. Transforming these complex data streams into actionable clinical and biological insights requires next-generation computational methods. Machine learning, deep neural networks, and generative AI have therefore become essential pillars of modern biomedical research.

The Bioinformatics and Omics Data Science Platform is a research group within the Berlin Institute of Medical Systems Biology (BIMSB) at the Max Delbrück Center (MDC). We pioneer machine learning frameworks and agentic AI systems for life science discovery, while investigating disease mechanisms in close collaboration with experimental and clinical partners across the MDC.

Unraveling complex disease

Our mission is to dissect the genetic, epigenetic, and regulatory mechanisms driving cell differentiation and complex diseases such as cancer. We build open-source tools for spatial biology and multi-omics integration (e.g., VoltRon, Flexynesis), alongside AI frameworks that leverage large language models (LLMs) and agentic workflows for clinical research and precision oncology (e.g., 2ndOpinion, Onconaut). Ultimately, we bridge genomics, multimodal health data, and autonomous data analysis to discover disease signatures, model therapeutic responses, and enable reproducible medical innovation.

The Bioinformatics and Omics Data Science Platform is a research group within the Berlin Institute of Medical Systems Biology (BIMSB) at the Max Delbrück Center (MDC). The platform creates and maintains bioinformatics tools and databases, and studies molecular mechanisms using externally available and/​or in-house computational techniques. In addition, the platform provides collaboration opportunities and support for MDC scientists.

Further information

Studying biological mechanisms with computational tools

We supply infrastructure and expertise for the bioinformatics and scientific IT needs of researchers, providing hardware and software for research-oriented tasks. We have various web-based interactive tools at our disposal, e.g. a local copy of the UCSC (University of California, Santa Cruz) Genome Browser, Shiny Server for interactive analysis, and an internal Galaxy server. We also have access to bioinformatics software solutions such as methylKit, genomation, RCAS, netSmooth, and a GNU Guix bioinformatics software repository.

We provide additional specialized IT services geared towards bioinformatics end users as well as a mobile teaching system for scientists. In addition, we offer courses and consultation sessions on bioinformatics, IT skills, and programming.

We provide additional specialized IT services geared towards bioinformatics end- users as well as a mobile teaching system for scientists. In addition, we offer courses and consultation sessions on bioinformatics, IT skills, and programming.

Databases & web apps

• Local copy of UCSC Genome Browser and track hubs to display genomics data
• Galaxy server for guided user interface for bioinformatics analysis
• Onconaut 
  • Description: An AI- and genomics-driven clinical decision-support research platform designed to navigate personalized cancer therapies. It bridges patient genomic variant profiles with clinical evidence bases (such as CIViC) to identify targeted therapy options, evaluate drug sensitivity vs. resistance markers, assess evidence quality tiers, and estimate targeted therapy eligibility across cancer types.
  • Links:
    • Onconaut Platform Web App

Software

• Shiny Server for interactive analysis
• methylKit
  • Description: An R/​Bioconductor package for DNA methylation analysis and annotation using high-throughput bisulfite sequencing data (RRBS, WGBS, target-capture, and base-pair 5hmC assays like TAB-seq or oxBS-seq). It provides differential methylation calling, clustering, and genomic annotation.
  • Links: GitHub Repository | Bioconductor Page
• genomation
  • Description: An R/​Bioconductor package for summarizing, annotating, and visualizing high-throughput genomic intervals (ChIP-seq peaks, methylation tracks, promoter profiles, and RNA-seq signals).
  • Links: GitHub Repository | Bioconductor Page
• RCAS (RNA Centric Annotation System)
  • Description: A tool for functional annotation and downstream analysis of custom RNA-target regions (e.g., CLIP-seq binding peaks, MeRIP-seq epitranscriptomic marks, or custom transcriptomic intervals). It outputs interactive dynamic reports with GO/gene-set enrichments and feature distribution plots.
  • Links: GitHub Repository | Bioconductor Page
• netSmooth
  • Description: An R/​Bioconductor package that uses network smoothing (diffusing gene expression signals along Protein-Protein Interaction networks) to reduce dropouts and noise in single-cell RNA-seq, significantly improving cell-type identification and clustering.
  • Links: GitHub Repository | Bioconductor Page
• VoltRon
  • Description: An R‑based spatial multi-omics analysis platform designed to handle diverse modalities, tile-based imaging, and varying spatial resolutions. It coordinates image registration, cell-cell communication, and spatial neighborhood analyses across multiple sequential or adjacent tissue layers.
  • Links: GitHub Repository | Documentation
• Flexynesis
  • Description: A Python-based deep-learning framework for integrating heterogeneous multi-omics bulk datasets (RNA-seq, methylation, proteomics, CNVs). It provides modular architectures (VAE, multi-task DNNs, GNNs) tuned for precision oncology tasks like subtype classification, drug response prediction, and survival analysis.
  • Links: GitHub Repository | PyPI Package
• ikarus
  • Description: A machine learning framework designed to distinguish tumor cells from normal cells at single-cell resolution. It extracts robust gene signatures by integrating reference datasets and refines malignant cell calls using cell-cell networks to boost classification accuracy across diverse tumor microenvironments.
  • Links: GitHub Repository | PyPI Package

Other Bioinformatics-related infrastructure services

• Reproducible bioinformatics software deployed at workstations and within the cluster via GNU Guix
• Virtual machines that can deploy publication-related web apps and databases such as DoRiNA and circBase
• compgenomr: Companion package and learning material for Dr. Akalin’s textbook ​“Computational Genomics with R”. Book Repository