Challenge: The client wanted to re-analyse a large number of published scRNA-seq datasets focused on chronic diseases to identify pathogenic cell types that could be targeted for treatment.

​

Solution: Jupyter notebook python scripts were developed for the entire workflow encompassing pre-processing & quality control, dimension reduction & batch correction, clustering and further subclustering of particular cell-types and identification of cell populations associated with phenotypes of interest. Custom differential gene expression tables were produced including annotation to aid filtering. Genes upregulated in pathogenic populations were identified and expression visualized in various plots. 

​

Impact: The pipeline was used by myself and other bioinformaticians to analyse a large number of datasets. This not only sped up analysis but ensured that results were obtained in a consistent manner, allowing direct comparison. Obtaining computational results quickly allowed hypotheses to be generated about what to test experimentally in the lab, moving towards the goal of generating treatments to target pathogenic diseases.