Researchers from St. Vincent’s Institute of Medical Research have introduced a new computational tool called Chloris that helps scientists better understand the complex structure of tumors at the single-cell level. Tumors are not uniform; they are composed of different groups of cells, called clones, each carrying distinct genetic changes known as copy number alterations, or CNAs. These variations can influence how tumor cells grow, respond to treatments, and evolve over time. Mapping these clonal structures accurately is critical for designing personalized cancer therapies and understanding why some tumors are resistant to certain treatments.

(A) An example process of tumor progression: starting from a copy-neutral normal cell, 5 distinct sub-clones evolve due to the accumulation of copy number gains and losses over time. (B) First row: The CN profile representation for each of the clones, where colors indicate 4 CN states: loss , neutral , gain1 , and gain2 , respectively. Second and third row: Stylized example of the associated input (relative depth ratio [RDR] and B-Allelic Fraction [BAF]) from a cell in the according clone, where RDR is loosely correlated with the copy number, while BAF provides an indication of the presence or absence of a CNA event. (C) Visual representation of the inputs: The RDR ⁠, which reflects the fluctuation in gene expression of a cell  relative to that of reference ⁠. The BAF (⁠⁠), visualized by their ratio ⁠.

Traditional approaches to studying CNAs often require researchers to know the number of clones in advance, and many methods treat cell clustering and CNA detection separately, which can reduce accuracy. Chloris overcomes these limitations by using a Bayesian modeling approach that integrates single-cell RNA sequencing data to simultaneously cluster cells into clones and identify CNAs within each clone. The model also incorporates information from germline single-nucleotide polymorphisms, allowing it to distinguish inherited genetic differences from tumor-specific alterations.

In practical applications, the team tested Chloris on data from human metastatic melanoma and anaplastic thyroid tumors. The tool successfully distinguished tumor cells from non-tumor cells, identified the correct number of clones, and mapped the CNAs within each clone. Importantly, the method revealed functional differences between clones in the same tumor, showing how variations in gene expression can relate to specific CNAs. By providing a detailed picture of tumor heterogeneity, Chloris can help researchers understand the progression of cancer, identify potential drug targets, and improve the design of precision therapies.

Chloris is available as an R package, making it accessible for computational biologists and cancer researchers who want to analyze single-cell RNA sequencing data with minimal manual intervention. Its automated and integrative approach represents a significant step forward in clonal analysis, offering a powerful tool to explore the diversity of cells within tumors and uncover insights that could guide future cancer treatments.

Availability – github.com/pqiao29/Chloris_experiment.