Here is a list of possible topics for Bachelor thesis in 2027 in our group. These are examples, and topics might slightly evolve depending on your precise interests, and the status of these projects at the beginning of the thesis!
Understanding how the loss or gain of a gene reshapes the transcriptional state of a cell is a central question in functional genomics. CRISPR knockout experiments give ground-truth answers, but are costly and slow. Computational tools that predict the effect of a perturbation from single-cell RNA-seq data have therefore become popular. They follow very different strategies:
However, these tools have rarely been compared against each other on the same data, so it is unclear which one to trust in a given context. In this project, the student will build a reproducible benchmarking pipeline and compare several of these tools against experimental single-cell perturbation data (e.g. Sciplex, Perturb-seq, CROP-seq), covering cancer cell lines, primary immune cells and stem cell differentiation models. The aim is to understand where each tool succeeds or fails, and whether its performance depends on the biological system.
RIG-I is a sensor of viral RNA that triggers antiviral and immune responses. Too little or too much RIG-I signalling can contribute to interferonopathies and other immune-mediated disorders, and recent studies have linked RIG-I-expressing cell subpopulations with antitumoral properties to improved patient outcomes. In previous work, we characterized a transcriptional signature of RIG-I signalling that is independent of the canonical interferon response. In this project, the student will use this signature to systematically map RIG-I expression and activation across healthy tissues and cell types (single-cell data from the Human Cell Atlas), across cancer types (TCGA), and across other human diseases (public transcriptomic resources). The goal is to rank tissues, cell types and diseases according to RIG-I activity, to compare cancer with healthy tissue, and to identify genes, pathways and gene regulatory programs associated with RIG-I activation.
After stimulation with viruses or synthetic dsRNA, only a subset of cells expresses the type I and type III interferons (IFN-β and IFN-λ), and the reasons for this heterogeneity are poorly understood. In our multimodal single-cell data of cells stimulated with dsRNA for 3 and 16 hours, we observed increased chromatin accessibility at regions associated with a RIG-I activation signature already at early time points, suggesting that epigenetic regulation contributes to the heterogeneous induction of interferons. In this project, the student will integrate scATAC-seq and scRNA-seq data, identify the cells that show the RIG-I signature, and stratify them by IFN-β/λ expression. Differential chromatin accessibility between interferon-expressing and non-expressing cells, together with transcription factor motif enrichment, will be used to identify candidate regulators of interferon expression.
Gene-expression signatures combine the expression of selected genes into a score that describes tumor biology or is associated with treatment outcome. For example, signatures reflecting blood-vessel formation have been linked to the outcome under drugs such as sunitinib in renal cell carcinoma (RCC), the most common type of kidney cancer. Different scoring methods can be applied to the same genes, but they may rank patients differently or behave differently across datasets. This matters when a signature is transferred between clinical studies, or from clear cell RCC to the less common papillary RCC. In this project, the student will compare established scoring approaches (mean expression, singscore and ssGSEA) on a small collection of signatures and curated kidney cancer datasets. The comparison will assess the robustness of the scores (e.g. to preprocessing, missing genes and sample number), their association with progression and survival, and their transferability between cohorts.