Overview: Wu et. al examined the causal mechanisms between noncoding loci and the variants causing disease by applying a new AI-based system, GrID-Net, that allows for deeper insight on the state between chromatin accessibility and gene expression (i.e., “cell state parallax”). The statistical concept of Granger causality, which states that event A can precede and predict event B, was the computational mechanism used to construct a network of locus-gene relationships across not only long distances but also with single-cell data. Snapshots of the cell state were analyzed between the epigenome and transcriptome within a single cell in an effort to construct causal rather than correlational associations between noncoding genes and the disease-causing variant. Unlike traditional methods, which formed connections based on gene proximity, results showed that GrID-Net was able to improve this ability by 36% locating 132 schizophrenia variant genes affected by noncoding loci. KCNG2 and SLC12A6, genes affected by noncoding loci, are vital transporters that enable potassium to reach the brain, thereby stimulating cognitive function and neural pathway connectivity. This could effectively decrease the expense and size of research experiments, given GrID-Net’s ability to analyze individual single-cell data and provide insight into diseases caused by noncoding loci.