Research
Interest and Overview
Our interests focus on the understanding of the molecular basis of gene regulation of diseases through translational research, specifically, but not exclusively, related to inflammation. The key aspects of our work include genome-wide gene expression analysis from human patient samples employing technologies such as RNA-Seq, Cap Analysis of Gene Expression (CAGE) or small RNA sequencing (miRNA). Our analysis goes beyond differentially expressed genes and identifies a variety of candidate elements responsible for the observed expression differences in the disease patients and the associated clinical phenotypes. Application of sequencing technology to the transcriptome previously has been utilized to uncover a range of regulatory elements and mechanisms, including regulation through transcription factors (TFs), nearby but distinct alternative promoters resulting in the same protein but employing different sets of regulatory TFs, expression of anti-sense RNA to modulate the sense-RNA and the regulatory role of expressed repeat elements and miRNAs. Subsequent functional validation studies confirm the suggested regulatory relationships.
Spatial Transcriptomics Projects
Spatial transcriptomics characterization of breast tumors
Our group is working on a cutting-edge spatial transcriptomics project focused on breast cancer. This research aims to characterize the molecular landscape of breast cancer by analyzing tissue architecture and gene expression patterns. By combining spatial transcriptomics data with image analysis and multiomics approaches, we aim to uncover novel insights into tumor biology, heterogeneity, and treatment responses. This integrative approach allows us to better understand breast cancer at a deeper level, paving the way for personalized therapeutic strategies.Currently, we are analysing ST dataset of 48 breast tissue (tumour and non-tumour)
Cellular Deconvolution of Breast Cancer Spatially Resolved Transcriptomics (SRT)
Our research focuses on studying the impact of various scRNA-seq datasets on cellular deconvolution of sequencing-based Spatially Resolved Transcriptomics (SRT), providing guidance on selecting appropriate references. We stratify breast cancer datasets by molecular and histological subtypes, starting with individual scRNA-seq samples and gradually increasing sample sizes. We also assess whether tailored or broader references improve deconvolution performance, including different subtypes and non-breast cancer tissues. Finally, we evaluate full scRNA-seq datasets from complete breast cancer studies to determine the need for feature engineering.
We use state-of-the-art deconvolution tools like Cell2Location, Tangram, and GraphST, assessing their performance through correlation analysis of cell type marker genes and proportions in low-resolution SRT datasets. Additionally, we compare deconvolution output to pseudo-spots with known cell type proportions from high-resolution datasets like Xenium.
Chromatin Accessibility integration for Breast Cancer Spatially Resolved Transcriptomics (SRT)
As a specific multi-omics approach implementation, our group aims to integrate the transcript information derived from our Spatial Transcriptomic breast cancer dataset with chromatin accessibility derived from ATAC-seq. In particular, we are using a Spatial ATAC-seq breast cancer dataset and, as a new integration approach, the linking point between the two techniques will be solely spot-level image similarity.
To perform this type of operation we’ll have to focus mainly on the image analysis steps, going through the implementation of picture preprocessing and deep-learning strategies to work on image normalisation and feature extraction. The final workflow will be able to stratify information in tissue regions and highlight the possible biological association between the two omic approaches.
DoGA project:
The project will generate the most comprehensive functional information source of the dog genome to facilitate the highest resolution disease gene mappings not possible with the current reference data. This new genomic resource will become publicly available and serve the international research community to better understand the molecular backgrounds of disease, morphology and behaviour for more efficient treatment scenarios. Our project will utilize the new genome annotation to identify risk variants for common brain disorders such as epilepsy, anxiety and neurodegeneration and to understand variable expression patterns in different parts of the dogs’ and wolves’ brains.
