Contest Setting

"Digital Source Cup" Million Single Cell Gene Expression Data Algorithm Contest: This contest aims to improve the performance of single-cell clustering algorithms, requiring participants to develop/optimize single-cell clustering algorithms to reduce memory and increase computing speed while ensuring the accuracy of analysis results using data resources provided by the organizer.
Single-cell transcriptome sequencing (scRNA-seq) has become a core tool for identifying and characterizing cell types, states, lineages, and networks. Cell clustering is a common step in the analysis of single-cell data. The goal of clustering analysis is to retain the major sources of variation in the dataset of the cell type to be defined while trying to shield variation due to useless sources of variation (sequencing depth, cell cycle differences, mitochondrial expression, batch effects, etc.). Several well-established single-cell clustering algorithms exist, such as classical hierarchical clustering and K-means, as well as SC3 clustering developed specifically for single-cell technology. However, as single-cell data continues to be produced, the throughput of data processing will continue to climb, so the performance of single-cell clustering algorithms needs to be continuously improved.