CeL-ID

CeL-ID authenticates cell lines by generating and comparing genomic variant profiles from RNA-seq data to detect misidentification and estimate cross-contamination against Cancer Cell Line Encyclopedia (CCLE) reference profiles.

Key Features:

• Variant Profile Utilization: Generates cell line–specific variant profiles from RNA-seq using an in-house pipeline, capturing allele frequencies, allelic fractions, and depth of coverage across genomic variants.
• Comprehensive Training and Testing: Trained and tested on over 900 RNA-seq datasets from the Cancer Cell Line Encyclopedia (CCLE) spanning adult and pediatric cancer cell lines.
• High Discriminatory Power: Performs pairwise comparisons of variant profiles in which distinct cell lines show significant differences, while identical, synonymous, and derivative lines exhibit high variant identity and highly correlated allelic fractions for classification.
• Contamination Estimation: Applies a linear mixture model to estimate potential cross-contamination and to identify contaminant CCLE cells when no perfect match is found.

Scientific Applications:

• Cell line authentication: Detects misidentified or synonymous cell lines in cancer biology studies using RNA-seq–derived variant signatures.
• Contamination detection: Estimates cross-contamination and identifies contaminant CCLE cell lines to support validity of experimental results and downstream analyses.
• Support for translational studies: Ensures correct cell line provenance for genomic analyses and evaluation of therapeutic agents.

Methodology:

Generates variant profiles from RNA-seq via an in-house pipeline, compares profiles pairwise using allele frequencies, allelic fractions, and depth of coverage, leverages high variant identity and allelic-fraction correlation for classification, and applies a linear mixture model to estimate cross-contamination; training and testing used >900 CCLE RNA-seq datasets.