DCI: Learning Causal Differences between Gene Regulatory Networks

Difference causal inference (DCI) identifies changes in gene regulatory relationships across two conditions directly from their expression data. Rather than reconstructing both complete networks, the algorithm estimates which causal edges appear, disappear, or change strength, reducing sample and computational requirements. A Python implementation includes stability selection across tuning parameters. Applications to single-cell RNA sequencing illustrate network comparison and validation through intervention-effect prediction.