PhyDOSE

PhyDOSE estimates the minimum number of single cells required for follow-up single-cell sequencing (SCS) experiments to distinguish among candidate tumor phylogenies derived from bulk DNA sequencing data at a specified confidence level.

Key Features:

• Input requirements: Accepts a set of candidate phylogenetic trees, a frequency matrix from bulk DNA sequencing data, and a user-specified confidence level.
• Optimization of cell number: Computes the minimum number of single cells required in a follow-up SCS experiment to confidently reconstruct the tumor phylogeny.
• Distinguishing features identification: Identifies unique distinguishing features among clones that differentiate one candidate tree from others.
• Probabilistic modeling: Uses a probabilistic model that incorporates clonal prevalence and candidate phylogenies to estimate sampling requirements.
• Error consideration: Accounts for typical errors associated with single-cell sequencing when inferring required sample sizes.

Scientific Applications:

• Cost efficiency: Reduces the number of cells required for follow-up SCS experiments, lowering sequencing and resource costs.
• High-fidelity phylogeny reconstruction: Improves accuracy in reconstructing tumor phylogenies, supporting analyses of tumor evolution and heterogeneity.
• Validation and use cases: Validated via simulations and retrospective analyses on leukemia patients and shown in prospective studies to disambiguate candidate trees by selecting cells across multiple biopsies, including complex cases such as lung cancer.

Methodology:

Employs a probabilistic model that incorporates distinguishing features from bulk data—notably clonal prevalence and candidate phylogenies—to infer the optimal number of single cells required while accounting for typical single-cell sequencing errors.