Annotate cells (legacy)

Uploading
Waiting for file…
Processing
Idle
This page conducts cell annotations on the uploaded gene expression files>
This online-page is optimized small number of cells. Limited performance for datasets containing more than thousands of cells. For better and faster performance, use pangeapy API.
  1. Input file configuration
    • Should contain gene expression matrix (cell_barcode × gene_id)
    • Raw expression must be 1e4-normalized & log1p-transformed
      normalized up to 10,000 counts per cell, then log-transformed with 1 pseudocount
    • File format: .csv or .csv.gz (recommended for faster performance)
  2. Cell annotation
    • Performed per cell barcode with a selectable model
    • PANGEA provides one Level1 model and ten Level2 models
    • Level1: 32 cell types; Level2 (combined): 165 annotations
    • Predictions are based on pre-trained logistic regression models
    • Results may differ from the original PANGEApy package (github.com/srkim727/pangeapy)
  3. Output file configuration
    • Output file: pred.csv with three columns for each cell barcode
    • predicted_label – predicted cell label from the selected model
    • conf_score – confidence score from the model prediction
    • cert_score – certainty compared to other labels within the model
Log

results matching ""

    No results matching ""