top of page

Alliance Interinstitutional Postdoc Program Positions

Two Alliance Interinstitutional Postdoc Program positions are currently open for applicants with co-supervision by Center SynGen Research Group Leader Alexander Sasse.

Application deadline: March 31, 2026, 5:00 pm CEST.
Please visit the Health + Life Science Alliance website for further information, application procedures, and other project outlines.
View these two project descriptions below.
Sasse/Furlong
Decoding gene regulation through integrated sequence-to-function modeling of genetic variation 
Sasse / UniHD & Furlong / EMBL 
Alexander Sasse
ZMBH, Center for Synthetic Genomics,
Universität Heidelberg, Heidelberg 

a.sasse@zmbh.uni-heidelberg.de

https://www.zmbh.uni-heidelberg.de/sasse/default.shtml

Eileen Furlong
Genome Biology, EMBL,
Heidelberg 
furlong@embl.de
https://furlonglab.embl.de

Project Outline 

Gene regulation during development requires precise coordination between cis-regulatory elements and the chromatin landscape. While the Sasse lab has developed advanced sequence-to-function models for predicting regulatory grammar, and the Furlong lab has mapped enhancer activity across developmental time, a critical gap remains in understanding how genetic variants dynamically influence enhancer function across developmental stages. This project will integrate deep learning approaches with developmental genomics to create predictive models that decode how sequence variants modulate enhancer activity throughout embryonic development, focusing on the temporal dimension of gene regulation and how sequence changes alter enhancer behavior across developmental transitions.

The postdoc will develop enhanced sequence-to-function models incorporating natural sequence variation using single cell data from the Furlong lab (scATAC-seq) in both wildtype and F1 embryos from different genetic background, using Drosophila embryonic development as a 'proof-of-principle' model system to predict both enhancer activity and the impact of genetic variation. By training models on both bulk and single cell chromatin accessibility, the project will create a frameworks that links sequence features to regulatory output. This work will produce interpretable models with testable predictions of the functional impact of natural sequence variation on transcription factor binding and enhancer activity. 
 

Required Qualifications 

The ideal candidate will have strong computational skills and demonstrated expertise in machine learning or statistical inference, with hands-on experience applying deep learning models to genomics datasets. Experience working with single-cell genomics data is a significant advantage. The candidate should excel at integrating computational analysis with biological interpretation and thrive in a collaborative, interdisciplinary research environment. 

They must possess outstanding verbal and written communication skills, strong analytical thinking, and the ability to independently design, execute, and interpret computational workflows. A clear commitment to scientific rigor, data quality, and reproducible research practices is essential. 

The project is not suitable for clinician scientists. 
 

Alexander Sasse - Research Group Description 

The Sasse lab’s research focuses on developing deep learning-based sequence-to-function (S2F) models to understand how regulation of gene expression is encoded by genomic sequences. His group works at the intersection of computational biology and synthetic genomics, using machine learning to analyze large-scale gene expression and cis-regulatory element (CRE) datasets. These models integrate diverse data types, including single-cell genomic data, to predict gene regulatory functions across different cell types and species. The goal is to decode the sequence grammar that controls context-dependent gene regulation and to understand how genetic variants impact cellular phenotypes and disease. 

Sasse's team also develops generative models for designing synthetic regulatory elements with tailored properties to advance biotechnology applications. Their research aims to uncover molecular mechanisms underlying gene regulation, evolution of cis-regulatory grammar, and to provide tools for interpreting disease-associated non-coding genetic variants. This work holds promise for synthetic biology, gene therapy, and improving understanding of gene regulation in health and disease. 

Relevant Publications 

  1. Refining sequence-to-activity models by increasing model resolution. bioRxiv [Preprint]. Doi: 10.1101/2025.01.24.634804 

  2. Deep Genomic Models of Allele-Specific Measurements. bioRxiv [Preprint]. Doi: 10.1101/2025.04.09.648060 

  3. Unlocking gene regulation with sequence-to-function models. doi: 10.1038/s41592-024-02331-5 

  4. Benchmarking of deep neural networks for predicting personal gene expression from DNA sequence highlights shortcomings. doi: 10.1038/s41588-023-01524-6 

  5. A resource of RNA-binding protein motifs across eukaryotes reveals evolutionary dynamics and gene-regulatory function. doi: 10.1038/s41587-025-02733-6 

​

Eileen Furlong - Research Group Description 

The Furlong Lab focuses on understanding the general principles of genome regulation during embryonic development. They study how the cis-regulatory genome is organized within the nucleus and how chromatin state and transcription factors influence gene expression patterns in multicellular embryos. Their research integrates single-cell genomics, genetics, high-resolution imaging, and computational biology to decode transcriptional regulation. A key area is exploring enhancer function during developmental transitions, including how multiple enhancers and cis-regulatory elements collaborate to regulate gene expression. They use

innovative methods such as CRISPR gene editing and optogenetics to manipulate transcription factors dynamically and dissect their regulatory roles. The lab also investigates 3D chromatin loops and their influence on gene regulation. 

The Furlong group employs large-scale single-cell assays, including scATAC-seq, to map active regulatory elements across developmental time in Drosophila embryos. They explore how genetic variation affects enhancer function and gene regulation, applying deep learning to predict the impact of DNA variants on transcription factor binding. Their work bridges developmental biology, genomics, and computational approaches to elucidate how transcriptional networks drive cell fate decisions and ensure robust developmental progression.

Relevant Publications

  1. Integrating genetic variation with deep learning provides context for variants impacting transcription factor binding during embryogenesis. doi: 10.1101/gr.279652.124 

  2. Targeted design of synthetic enhancers for selected tissues in the Drosophila embryo. doi: 10.1038/s41586-023-06905-9 

  3. scDALI: modeling allelic heterogeneity in single cells reveals context-specific genetic regulation. doi: 10.1186/s13059-021-02593-8 

  4. Simultaneous cellular and molecular phenotyping of embryonic mutants using single-cell regulatory trajectories. doi: 10.1016/j.devcel.2022.01.016 

  5. The continuum of Drosophila embryonic development at single-cell resolution. doi: 10.1126/science.abn5800

Dieterich/Sasse
Interpretable Generative Deep Learning Models for improved Nanopore RNA signal processing to robustly detect RNA modifications and structure
Dieterich / Med Fak HD & Sasse / UniHD 
Alexander Sasse
ZMBH, Center for Synthetic Genomics,
Universität Heidelberg, Heidelberg 

a.sasse@zmbh.uni-heidelberg.de

https://www.zmbh.uni-heidelberg.de/sasse/default.shtml

Christoph Dieterich 
Department of Internal Medicine III,
University Hospital Heidelberg, Heidelberg  
christoph.dieterich@med.uni-heidelberg.de 
www.dieterichlab.org 

Project Outline 

Nanopore Direct RNA Sequencing (DRS) measures RNA molecules as continuous, analog time-series signals generated when individual nucleotides pass through a protein nanopore under an applied voltage. Instead of detecting bases via fluorescence or synthesis, DRS captures perturbations in ionic current caused by an RNA substring (typically few nt) occupying the pore’s constriction, turning each RNA molecule into a noisy physical sensor readout. Because RNA is sequenced natively, without reverse transcription or amplification, the raw signal encodes a superposition of sequence, secondary structure, chemical modifications, and translocation kinetics. This project will develop generative AI models that map between Nanopore direct-RNA sequencing (DRS) time-series and molecular sequence, and annotations describing RNA modifications and secondary structure. Using existing and novel DRS traces from designed pools of mRNA and tRNA sequences with and without characteristic chemical modification patterns and secondary structures, we train generative multi-modal models to (a) predict modifications/structure from signal (forward model) and (b) use model interpretation methods to understand the time series signals required for accurate predictions (inverse problem). This combined experimental–computational loop will reveal how modifications and local RNA structure shape the ionic current signal and enable better modification calling, secondary structure prediction and signal simulation. 
 

Required Qualifications 

Essential Skills

  • Background in computer science, statistics, bioinformatics, physics or similar. 

  • Strong programming skills for data analysis, statistics, and visualization. 

  • Experience with Linux-based environments, version control (Git), and reproducible workflows. 

  • Ability to shape and build deep learning architectures. 

Preferred / Highly Advantageous Skills 

  • Previous work in the context of RNA biology. 

  • Familiarity with Nanopore technology. 

  • Hands-on experience with RNA sequence folding / design. 

​Professional Competencies 

  • Strong analytical thinking and capacity to independently design and interpret computational workflows. 

  • Strong motivation to work at the intersection of experimental and computational biology, with close interaction with in vivo research teams. 

  • Outstanding verbal and written communication skills to support effective interdisciplinary collaboration. 

  • A clear commitment to scientific rigor, data quality, and reproducible research practices. 

The project is not suitable for clinician scientists. 

​

Christoph Dieterich - Research Group Description

The Dieterich Lab bridges RNA biology, systems cardiology, and clinical data science. Our research focuses on three major areas: 

First, we investigate RNA maturation and processing, and our lab has developed a broad suite of software tools that enable the exploration of the complex RNA landscape, particularly using long-read nanopore sequencing and AI-supported analysis of RNA modifications, splicing and translation. Second, we have established a strong focus on systems cardiology, using both in vitro and in vivo models of heart failure to uncover transcriptomic, translational, and molecular mechanisms underlying cardiac dysfunction. 

Third, through our involvement in the HiGHmed Consortium within the German Medical Informatics Initiative, we connect computational biology with clinical data science, enabling secure, interoperable, and analysis-ready datasets for translational research. 

Across all these domains, we actively develop and apply machine learning methods, including sequence models and large language models, to extract meaningful biological and clinical insights. Our overarching mission is to merge computational innovation with experimental and clinical research, advancing data-driven science in cardiology.

Relevant Publications

  1. International Human RNome Project Consortium. Unlocking the regulatory code of RNA: launching the Human RNome Project. Genome Biol. 2025 Oct 24;26(1):367. doi: 10.1186/s13059-025-03824-y. 

  2. Chan A, Naarmann-de Vries IS, Dieterich C. Ψ-co-mAFiA: concurrent detection of pseudouridine and m6A in single RNA molecules. Bioinformatics. 2025 Oct 2;41(10):btaf536. doi: 10.1093/bioinformatics/ 

  3. Rabolli C, Longenecker JZ, Naarmann-de Vries IS, Serrano J, Petrosino JM, Kyriazis GA, Dieterich C, Accornero F. The cardiac METTL3/m6A pathway regulates the systemic response to Western diet. JCI Insight. 2025 Apr 24;10(11):e188414. doi: 10.1172/jci.insight.188414. 

  4. Boileau E, Wilhelmi H, Busch A, Cappannini A, Hildebrand A, Bujnicki JM, Dieterich C. Sci-ModoM: a quantitative database of transcriptome-wide high-throughput RNA modification sites. Nucleic Acids Res. 2025 Jan 6;53(D1):D310-D317. doi: 10.1093/nar/gkae972. 

  5. Chan A, Naarmann-de Vries IS, Scheitl CPM, Höbartner C, Dieterich C. Detecting m6A at single-molecular resolution via direct RNA sequencing and realistic training data. Nat Commun. 2024 Apr 18;15(1):3323. doi: 10.1038/s41467-024-47661-2. 

​

Alexander Sasse - Research Group Description 

The Sasse Lab develops Deep Learning based Sequence-to-function models to decode the gene regulatory syntax of eukaryotic cells. By integrating multimodal genomic data, the lab builds models that relate regulatory DNA and RNA sequence to cellular phenotypes, enabling the identification of regulatory elements and factors, the prediction of cis-regulatory variant effects, and the mechanistic interpretation of gene expression control. These tools support both fundamental and quantifative insight into how cis-regulatory elements encode gene expression and practical applications such as designing synthetic regulatory sequences and informing therapeutic strategies, including gene and mRNA-based treatments.  

Relevant Publications

  1. Sasse, A.*, Ng, B.*, Spiro, A.E.*, et al. 2023, Benchmarking of deep neural networks for predicting personal gene expression from DNA sequence highlights shortcomings. Nature Genetics 55, 2060–2064, https://doi.org/10.1038/s41588-023-01524-6 

  2. Sasse, A., Chikina, M. & Mostafavi, S. Unlocking gene regulation with sequence-to-function models. 2024, Nature Methods Comment. https://doi.org/10.1038/s41592-024-02331-5 

  3. Chandra, NA., Hu Y., Buenrostro JD., Mostafavi S., Sasse A., 2025. Refining the Cis-Regulatory Grammar Learned by Sequence-to-Activity Models by Increasing Model Resolution. bioRxiv, https://doi.org/10.1101/2025.01.24.634804 (submitted to Bioinformatic Advances) 

  4. Tu, X.*, Sasse A.*, Chowdhary K.*, Spiro AE., Yang L., Chikina M., Benoist CO., Mostafavi S., 2025. Deep Genomic Models of Allele-Specific Measurements. bioRxiv, https://doi.org/10.1101/2025.04.09.648060 (submitted to RECOMB proceedings) 

  5. Sasse, A.*, Ray D.*, Laverty KU.*, Tam CL., Albu M., et al. 2025, A resource of RNA-binding protein motifs across eukaryotes reveals evolutionary dynamics and gene-regulatory function. Nature Biotech., https://doi.org/10.1038/s41587-025-02733-6 â€‹

bottom of page