🤖 AI Summary
Traditional feature selection methods (e.g., LASSO) struggle to identify genes associated with cell-state transitions in scRNA-seq data due to its high dimensionality, nonlinearity, and multicollinearity. To address this, we introduce quantum annealing—applied here for the first time to single-cell feature selection—by formulating a Quadratic Unconstrained Binary Optimization (QUBO) model that optimizes nonlinear gene subsets to capture dynamic expression patterns significantly correlated with differentiation trajectories. Our method integrates D-Wave quantum hardware for QUBO solving, standardized scRNA-seq preprocessing, and differential expression analysis. Evaluated on human cell differentiation datasets, it successfully recovers multiple known nonlinearly regulated differentiation genes, achieves a 12.3% improvement in classification accuracy, and substantially enhances biological interpretability. This approach overcomes the representational limitations of linear models in capturing complex regulatory relationships underlying cellular dynamics.
📝 Abstract
Feature selection is vital for identifying relevant variables in classification and regression models, especially in single-cell RNA sequencing (scRNA-seq) data analysis. Traditional methods like LASSO often struggle with the nonlinearities and multicollinearities in scRNA-seq data due to complex gene expression and extensive gene interactions. Quantum annealing, a form of quantum computing, offers a promising solution. In this study, we apply quantum annealing-empowered quadratic unconstrained binary optimization (QUBO) for feature selection in scRNA-seq data. Using data from a human cell differentiation system, we show that QUBO identifies genes with nonlinear expression patterns related to differentiation time, many of which play roles in the differentiation process. In contrast, LASSO tends to select genes with more linear expression changes. Our findings suggest that the QUBO method, powered by quantum annealing, can reveal complex gene expression patterns that traditional methods might overlook, enhancing scRNA-seq data analysis and interpretation.