Researchers from the University of Missouri and the Ohio State University say they have created a new way to analyze data from single-cell RNA-sequencing by using machine learning. The method uses the power of computers to intelligently analyze large amounts of data and help scientists draw faster conclusions and move to the next stage of the research.
Their study “scGNN is a novel graph neural network framework for single-cell RNA-Seq analyses” appears in Nature Communications.
“Single-cell RNA-sequencing (scRNA-Seq) is widely used to reveal the heterogeneity and dynamics of tissues, organisms, and complex diseases, but its analyses still suffer from multiple grand challenges, including the sequencing sparsity and complex differential patterns in gene expression,” write the investigators.
“We introduce the scGNN (single-cell graph neural network) to provide a hypothesis-free deep learning framework for scRNA-Seq analyses. This framework formulates and aggregates cell–cell relationships with graph neural networks and models heterogeneous gene expression patterns using a left-truncated mixture Gaussian model. scGNN integrates three iterative multi-modal autoencoders and outperforms existing tools for gene imputation and cell clustering on four benchmark scRNA-Seq datasets.”
“In an Alzheimer’s disease study with 13,214 single nuclei from postmortem brain tissues, scGNN successfully illustrated disease-related neural development and the differential mechanism. scGNN provides an effective representation of gene expression and cell–cell relationships. It is also a powerful framework that can be applied to general scRNA-Seq analyses.”
“Single cell genetic profiling is on the cutting edge of today’s technological advances because it measures how many genes are present and how they are expressed from the level of an individual biological cell,” said Dong Xu, PhD, a professor in the MU College of Engineering. “At a minimum, there could be tens of thousands of cells being analyzed in this manner, so there ends up being a huge amount of data collected. Currently, determining conclusions from this type of data can be challenging because a lot of data must be filtered through in order to find what researchers are looking for. So, we applied one of the newest machine-learning methods to tackle this problem—a graph neural network.”
After computers intelligently analyze the data through a machine learning process, the graph neural network then takes the results and creates a visual representation of the data to help easily identify patterns. The graph is made up of dots, each representative of a cell, and similar types of cells are color coded for easy recognition. Xu said precision medicine is a good example of how single-cell RNA-sequencing can be used.
“With this data, scientists can study the interactions between cells within the micro-environment of a cancerous tissue, or watch the T-cells, B-cells and immune cells all try to attack the cancerous cells,” Xu added. “Therefore, in cases where a person has a strong immune system, and the cancer hasn’t fully developed yet, we can learn how the cancer can possibly be killed at an early stage, and we have our results sooner because of machine learning, which leads us to a viable treatment faster.”
Xu believes this is a great example of how engineers and biologists can work together to study problems or issues in biology. He hopes this method can be used by biologists as a new tool to help solve complex biological questions, such as a possible treatment for Alzheimer’s disease.