Figure 3 .

Three graph-based learning models for single-cell multi-omics sequencing. All of them have leveraged on the neighborhood relationship for data imputation, data integration and data pairs correction.