During my PhD research, I developed the Zipper plot, a tool that can be used to assess the reliability of the annotation of thousands of transcripts using features that are indicative of independent transcription and that has been used to refine the human transcriptome generated using the RNA Atlas dataset. Furthermore, I evaluated the use of different RNA fractions (other than mRNAs) in the computational deconvolution of transcriptomics data and performed a comprehensive assessment of different key factors affecting the deconvolution results, including data transformation, scaling/normalization and marker selection strategies, the cell type composition of the reference matrix and the choice of method.

The next goal will be the establishment of a bioinformatics pipeline for deconvolution of the circulating tumor transcriptome (e.g. from blood or other biofluids) for diagnosis, follow-up and potential drug target identification for cancer patients.