Hi, I’m Irem 👋

Hi, I’m Irem 👋

PhD candidate in computational epigenomics and machine learning at Saarland University (ICBB Lab), working where machine learning meets gene regulation.

I build interpretable, biologically informed models that help explain how cells with identical DNA establish and change their identity, with a focus on graph neural networks that embed real biological structure (such as transcription-factor interaction networks) directly into how they learn. My research spans single-cell and bulk multi-omics, explainable AI, and causal inference, and has contributed to studies of the human immune epigenome (Nature Genetics, 2025).

🔬 Interests: explainable ML, graph neural networks, gene regulation, multi-omics, causal inference, open-source bioinformatics

🛠️ Open-source software

  • deconvR — R/Bioconductor package for deconvolution of bulk omics profiles
  • methylTFR — transcription-factor activity from DNA methylation
  • ChrAccR — unified single-cell and bulk ATAC-seq analysis
  • ARTEMIS — attention GNN with transcription-factor interaction priors (in preparation)

🌐 More: igunduz.github.io · ORCID

💬 If you’re working on AI for biology or drug discovery, I’d be glad to connect.