New tool automates cell identification in complex datasets

An international research group led by The University of Osaka has developed scODIN, a novel computational tool to classify cell types from single-cell RNA sequencing (scRNA-seq) data. Existing methods struggle to balance speed and accuracy, often misclassifying rare or transitional cells. scODIN overcomes this limitation by combining a hierarchical classification system (Tier system) with k-nearest neighbor inference. This approach allows for the rapid and accurate classification of large datasets, processing 650,000 cells in just six minutes. The tool’s improved accuracy stems from its ability to identify cells at varying levels of detail, recognize intermediate phenotypes through double labeling, and recover cells affected by dropout events. scODIN promises to accelerate biomedical discoveries by enabling more precise and efficient analysis of complex biological processes and disease mechanisms.Read More