This is a project which is currently making use of HPC facilities at Newcastle University. It is active.
For further information about this project, please contact:
This project uses single-nucleus RNA sequencing (snRNA-seq) to study how donated human kidneys respond to machine perfusion, a technique in which fluid is continuously pumped through an organ outside the body to preserve it and assess its quality before transplantation. The work uses a novel perfusion system developed at Newcastle University and the Freeman Hospital. By profiling gene expression in thousands of individual cells from kidney tissue sampled during perfusion, we aim to map how different kidney cell types (including epithelial, endothelial, stromal and immune cells) respond to perfusion and to any therapies delivered through it, and to identify cellular signatures of injury, repair and organ viability. The HPC facility will be used to process raw sequencing data (alignment and quantification), carry out quality control, integrate data across samples, and perform downstream analyses such as cell-type annotation, differential gene expression and cell–cell communication inference. The long-term aim is to improve how donor kidneys are preserved and assessed, so that more organs can be safely transplanted.
Raw sequencing data will be processed using 10x Genomics Cell Ranger (including intronic reads, as is standard for single-nucleus data) or STARsolo for alignment to the human reference genome and gene-expression quantification. Quality control will include ambient RNA removal (CellBender) and doublet detection (Scrublet or scDblFinder). Downstream analysis will be carried out mainly in Python using Scanpy and the AnnData framework, with dataset integration using scVI-tools and/or Harmony, automated and manual cell-type annotation (e.g. CellTypist), differential expression and pathway analysis (e.g. decoupler), and cell–cell communication inference (e.g. CellPhoneDB or LIANA). Some analyses may also use R/Bioconductor packages. Software will be managed with conda environments and/or Singularity/Apptainer containers, and run as batch jobs through the scheduler as well as interactively via Jupyter. Processing is mainly CPU- and memory-intensive (alignment and large in-memory matrices), with occasional GPU use for deep-learning-based tools such as CellBender and scVI.