scESI: evolutionary sparse imputation for single-cell transcriptomes from nearest neighbor cells

Abstract

The ubiquitous dropout problem in single-cell RNA sequencing technology causes a large amount of data noise in the gene expression profile. For this reason, we propose an evolutionary sparse imputation (ESI) algorithm for single-cell transcriptomes, which constructs a sparse representation model based on gene regulation relationships between cells. To solve this model, we design an optimization framework based on nondominated sorting genetics. This framework takes into account the topological relationship between cells and the variety of gene expression to iteratively search the global optimal solution, thereby learning the Pareto optimal cell–cell affinity matrix. Finally, we use the learned sparse relationship model between cells to improve data quality and reduce data noise. In simulated datasets, scESI performed significantly better than benchmark methods with various metrics. By applying scESI to real scRNA-seq datasets, we discovered scESI can not only further classify the cell types and separate cells in visualization successfully but also improve the performance in reconstructing trajectories differentiation and identifying differentially expressed genes. In addition, scESI successfully recovered the expression trends of marker genes in stem cell differentiation and can discover new cell types and putative pathways regulating biological processes.

Multi-Objective Genetic Algorithm for Cluster Analysis of Single-Cell Transcriptomes

“…Interestingly, applications of MOO in single-cell transcriptomics are limited and mostly used to find the best combination of hyperparameters for deep learning models [ 62 ] and to impute missing values [ 63 , 64 ]. For instance, dropouts in single-cell transcriptomics are notoriously difficult to resolve.…”

“…When unresolved, they propagate noise to downstream analyses and lead to low-quality clustering results. By taking into account topological relationships between cells, the construction of cell–cell affinity matrices can be formulated as a MOO problem [ 64 ]. The structure learned from these matrices reduces gene expression noise and improves downstream results.…”

Deep Learning and Machine Learning Applications in Biomedicine

“…In scRNA-seq, a low RNA capture rate frequently leads to dropout issues. Researchers utilize neural network algorithms for data imputation in scRNA-seq, effectively mitigating noise in gene expression profiles [20][21][22]. It is noteworthy that a significant advantage of DL in scRNA-seq data analysis is its capacity to handle nonlinear relationships between genes.…”