Welcome to Celligrate!
Celligrate is a project for cell type characterization and integration from single cell RNA-sequencing (scRNA-seq) data. The backbone of Celligrate consists of two carefully-designed and extensively-validated computational algorithms: NS-Forest and FR-Match. NS-Forest is a random forest machine learning algorithm for cell type marker gene identification. FR-Match is a topological graph theory-based statistical learning algorithm for cell type matching. Celligrate also introduces a notion of “cell type barcode” for insightful visualization of cell type expression data. The use of Celligrate extends the utility of the upstream scRNA-seq analysis pipelines to downstream use cases, and ultimately accelerates the growth of knowledge about cell types by pooling results from individual studies.
See Methods for more details.
Software Suite
Tutorials
NS-Forest
FR-Match
- Tutorial 1: FR-Match data object + matching across overlapping neuroanatomical regions (Layer 1 and full MTG)
- Tutorial 2: Matching across non-overlapping neuroanatomical regions (M1 and MTG)
- Tutorial 3: Matching mouse M1 cell subclasses across different sample types (scRNAseq and snRNAseq)
- Tutorial 4: Matching human M1 cell subclasses across different platforms (SMART-seq and 10X)
- Tutorial 5: Cell type calling for spatial transcriptomics data (smFISH and scRNAseq)
- Tutorial 6: Notebook for challange 1 of Allen Institute Data challenge: Mapping of Cell Type Data
Demo data in the tutorials
The demo data storage is under migration. For any inquiries, please contact zhangy@jcvi.org.
News
-
FR-Match v2.0 paper is published
The FR-Match v2.0 paper is published in Scientific Reports!
here -
NS-Forest v2.0 paper is published
NS-Forest v2.0 paper is published in Genome Research!
here. -
FR-Match v1.0 paper is published
The FR-Match method paper is published in Briefings in Bioinformatics!
here -
NS-Forest v1.3 stable release
NS-Forest v1.3 is a stable release on GitHub here. More background information in this post.