Single-cell Solutions for Mining New Plant Marker Genes

In multicellular organisms, cell types and cell-specific functions arise largely from the differential expression of different genes in the cell. Recently, single-cell RNA sequencing (scRNA-seq) has emerged as a powerful method for studying gene expression in multicellular organisms. As a translational technique, scRNA-seq is particularly important for plant research because traditional methods for determining gene expression in individual cell types rely on transgenic lines expressing cell type-specific fluorescent markers, which are unavailable in most non-model species. Determining cell types is a critical step in analyzing and interpreting scRNA-seq data, primarily through differential analysis to identify genes characteristic of a cell subpopulation, combined with marker genes to identify cell types. Identifying new marker genes is particularly important for plant biology studies, as cell-type markers in non-model species are largely unknown.

Fig.1. Identification of marker genes with three different ways to label cells in Arabidopsis root. (Yan H, et al., 2022)

What We Offer

An essential step in analyzing single-cell RNA sequencing data is the use of marker genes to classify cells into specific cell types. Several computational methods have been developed to identify novel marker genes from scRNA-seq data in non-plant systems; however, none of these methods have been applied to plant systems. Our technical team is dedicated to developing and comparing machine-learning based methods to identify marker genes by analyzing feature importance.

Lifeasible provides professional single-cell solutions for mining new plant marker genes to identify highly expressed marker genes in each cell population, further facilitating cell type identification in scRNA-seq data from different plant species. Our new marker gene mining service helps to elucidate cellular heterogeneity in-depth and is critical for identifying cell populations of unknown cell types during plant development.

We selected the best-performing methods, Random Forest (RF) and SVM, to identify marker genes. These new markers not only assign cell types consistently with previously known cellular markers but also outperform existing markers in several evaluation metrics, including accuracy and sensitivity.

Advantages of Machine-Learning Based Marker Detection:

While most traditional methods select one gene at a time by calculating gene specificity, machine-learning methods characterize marker genes by analyzing combinations of many marker genes.
Machine-learning methods provide many principled ways to evaluate marker performance, allowing us to compare different marker genes more rigorously and unbiasedly.
Markers with novel expression patterns can be identified.
Assign cell types based on cells labeled using published methods.
Assigning cell types based on cells labeled with published methods.
Assign cell types based on internal GFP markers.

Lifeasible offers cutting-edge machine-learning methods to identify cell type marker genes from plant scRNA-seq data. We are also exploring cross-species mapping of plant scRNA-seq data based on marker genes and their expression patterns in roots to define root cell types for non-model species. If you are interested in our services or have some questions, please feel free to contact us or make an online inquiry.

Reference

Yan H, Lee J, Song Q, et al. Identification of new marker genes from plant single-cell RNA-seq data using interpretable machine learning methods[J]. New Phytologist, 2022, 234(4): 1507-1520.

The services provided by Lifeasible cover all aspects of plant research, please contact us to find out how we can help you achieve the next research breakthrough.

Contact

*If your organization requires the signing of a confidentiality agreement, please contact us by email.

For research use only, not intended for any clinical use.

Related Services