scACCorDiON: A clustering approach for explainable patient level cell cell communication graph analysis

Motivation The combination of single-cell sequencing with ligand-receptor analysis paves the way for the characterization of cell communication events in complex tissues. In particular, directed weighted graphs stand out as a natural representation of cell-cell communication events. However, current computational methods cannot analyze sample-specific cell-cell communication events, as measured in single-cell data produced in large patient cohorts. Cohort-based cell-cell communication analysis presents many challenges, such as the non-linear nature of cell-cell communication and the high variability presented by the patient-specific single-cell RNAseq datasets.

Here, we present scACCorDiON (single-cell Analysis of Cell-Cell Communication in Disease clusters using Optimal transport in Directed Networks), an optimal transport algorithm exploring node distances on the Markov Chain as the ground metric between directed weighted graphs. Additionally, we derive a k-barycenter algorithm using the Wasserstein-based distance, which is able to cluster directed weighted graphs. We compare our approach with competing methods in several large cohorts of scRNA-seq data. Our results show that scACCorDiON can predict clusters better, matching the disease status of samples. Moreover, we show that barycenters provide a robust and explainable representation of cell cell communication events related to the detected clusters. We also provide a case study of pancreas adenocarcinoma, where scACCorDion detects a sub-cluster of disease samples associated with changes in the tumor microenvironment. Availability The code of scACCorDiON is available at https://scaccordion.readthedocs.io/en/latest Contact ivan.costa{at}rwth-aachen.de Competing Interest Statement The authors have declared no competing interest.