Emerging single-cell technologies offer detailed cellular atlases of diverse tissues. However, the complexity of single-cell sequencing data poses challenges for effective analysis. Dr Vafaee and her team at UNSW are developing innovative approaches to overcome these challenges and improve the analysis of scRNA-seq data. During the talk, Dr Vafaee will highlight some of the key methods developed by her lab. One such method is a novel feature-extraction technique inspired by power spectral density analysis. This approach significantly enhances the accuracy of cell subtype separation in large-scale single-cell omics data. Additionally, she will present their "omic imagification" method, which transforms molecular measurements into two-dimensional RGB images. Leveraging convolutional neural networks, this method improves the classification of different biological phenotypes. Supervised cell type classification methods using external annotated source data gain popularity but often face limitations in accurately classifying cell types absent in the source data. Dr Vafaee will also briefly introduce her lab’s research on open-set domain adaptation techniques, allowing cell type annotation in target datasets using source datasets with non-overlapping cell types.