Review of Machine Learning Advancements for Single-Cell Analysis

Massive amounts of data describing the genomic, transcriptomic, and epigenomic features of many different cells are produced by single-cell omics approaches. Artificial intelli- gence (AI) models are widely used to infer biological informa- tion and construct predictive models from these data due to their adaptability, scalability, and remarkable success in other disciplines. Low-dimensional representations of single-cell omics data, batch normalization, cell type classification, trajectory inference, gene regulatory network inference, and multimodal data integration have all seen a recent surge in innovation thanks to machine and deep learning. We provide a survey of recent developments in machine learning (ML) algorithms intended for analysis of single-cell omics data to aid readers in navigating this rapidly-evolving literature.