Comparing the genomes of cutaneous melanoma tumors to commercially available cell lines
Metrics: PDF 1616 views | HTML 3559 views | ?
Stephen A. Luebker1, Weiwei Zhang1 and Scott A. Koepsell1
1Department of Pathology and Microbiology, University of Nebraska Medical Center, Omaha, NE
Stephen A. Luebker, email: [email protected]
Keywords: pre-clinical; tumor model; sequencing
Received: September 25, 2017 Accepted: November 12, 2017 Published: December 04, 2017
Insulated culture environment and prolonged propagation contribute to known limitations of cell lines, and selection is often limited to availability or favorable growth characteristics. To better characterize and improve selection of cell lines, we compared 60 melanoma cell lines profiled by the Cancer Cell Line Encyclopedia and 472 cutaneous melanoma tumors profiled by The Cancer Genome Atlas by DNA sequence and copy number alterations. All samples were scored for stromal and immune cell composition by the ESTIMATE algorithm, and 412 tumors with ≥ 60% tumor cell fraction were compared to cell lines. Uncharacterized early passage cell lines that lacked BRAF, NRAS, or NF1 mutations had near zero mean Pearson correlation of copy number alterations per gene to tumors and also tended to have higher stromal scores. The Comet Exact Test was applied to tumors and cell lines identifying three pairs of genes mutated in a mutually exclusive pattern in tumors but not cell lines: BRAF and NRAS, BRAF and NF1, as well as NRAS and PTEN. Additionally, 31 genes were more frequently mutated in cell lines than tumors. Avoiding cell lines with co-occurring mutually exclusive mutations and the fewest differentially mutated genes within a known distribution of genetic similarity to tumors by copy number alterations may optimize selection.
All site content, except where otherwise noted, is licensed under a Creative Commons Attribution 4.0 License.