PANGEA

PANGEA (Public ANotated Gene Expression Atlas) is a cell annotation framework based on a single-cell transcriptome database.
PANGEA framework includes

1. Database curation
2. Construction of reference cell atlases and cell annotation models
3. Meta analyses based on integrated cell annotations

I. Database curation

PANGEA database mainly consists of re-aligned datasets from public repositories (NCBI).

• This re-aligned database encompasses
  • 2,058 human in vivo samples (1,426 from non-malignant, and 632 from maligant donors)
  • 131 disease contexts
  • 45 organ identities
    
    
• Benefits of this re-aligned database includes
  • minimized computational biases
  • consistent cell QC standards
  • facilitate implementation of count-based models (SCVI, …)

II. Reference cell atlases

Integration of expression profiles in PANGEA database identified

1. 32 different major cell types (Level1)
2. 165 different cellular subtypes (Level2)

Gene expression patterns of these annotations can be explored here.
Organ distribution patterns of these annotations can be explored here.

III. Cell annotation models

Based on reference cell atlases, annotation models were built using CellTypist.
Full version of PANGEA cell annotation framework is available via python package, PANGEApy.
Web-app implementation of PANGEA cell annotation model is available here.

Repositories

1. PANGEApy: cell annotation package/pipeline
2. representative cell atlases
3. cell annotation models