IMGG: Integrating Multiple Single-Cell Datasets through Connected Graphs and Generative Adversarial Networks

Abstract

There is a strong need to eliminate batch-specific differences when integrating single-cell RNA-sequencing (scRNA-seq) datasets generated under different experimental conditions for downstream task analysis. Existing batch correction methods usually transform different batches of cells into one preselected “anchor” batch or a low-dimensional embedding space, and cannot take full advantage of useful information from multiple sources. We present a novel framework, called IMGG, i.e., integrating multiple single-cell datasets through connected graphs and generative adversarial networks (GAN) to eliminate nonbiological differences between different batches. Compared with current methods, IMGG shows excellent performance on a variety of evaluation metrics, and the IMGG-corrected gene expression data incorporate features from multiple batches, allowing for downstream tasks such as differential gene expression analysis.

Integrating Multiple Single-Cell RNA Sequencing Datasets Using Adversarial Autoencoders

“…The cells on the connected graph are highly similar, and the spatial distance is smaller than other cells of the same type across batches. Note that steps (5) and (6) are the procedures for constructing the cross-batch similarity cell connected graph in the IMGG model [ 29 ], which are optional in the IMAAE model. The difference is that if these steps are not used, random selection will be adopted when establishing mappings between different batches of cells of the same type in subsequent steps.…”

Recent Advances in Generative Adversarial Networks for Gene Expression Data: A Comprehensive Review

“…Wang et al [48] Developed IMGG to integrate multiple scRNA-seq datasets, enhancing downstream tasks like differential gene expression analysis.…”

“…In the context of integrating multiple scRNA-seq datasets, IMGG [48] and ResPAN [55] have shown remarkable efficacy. These models use GANs to eliminate non-biological differences between batches while preserving biological information.…”

PETrans: De Novo Drug Design with Protein-Specific Encoding Based on Transfer Learning

“…Traditional methods such as high-throughput screening are inefficient because the number of resources required is not balanced by the small number of hit compounds. Conventionally, the identification of promising lead structures is achieved by experimental high-throughput screening (HTS), but this is time-consuming and expensive [ 1 , 2 , 3 ]. A typical drug discovery cycle takes approximately 14 years [ 4 ] and costs approximately 800 million dollars [ 5 ].…”