celluloid

celluloid clusters single-cell sequencing (SCS) mutation data to reduce dataset complexity and enable more tractable phylogenetic inference in tumor evolution studies.

Key Features:

• Categorical data clustering: Clusters categorical vectors or matrices representing mutations characterized by presence, absence, and uncertainty in sequenced cells.
• Centroid-based aggregation: Organizes clusters around centroids to reduce the number of mutation instances for downstream analysis.
• High-precision grouping: Prioritizes precision to avoid erroneously pairing unrelated mutations in the reduced dataset.
• Scalability and runtime reduction: Reduces dataset size to lower runtime and raise the upper bound on analyzable SCS instance size.
• Phylogenetic preprocessing: Produces reduced datasets intended for input to phylogenetic inference methods.

Scientific Applications:

• Tumor phylogenetics: Facilitates construction of evolutionary trees (phylogenies) from cancerous mutations in SCS data.
• Large-scale SCS analysis: Enables analysis of larger single-cell sequencing datasets in cancer genomics and evolutionary biology by reducing computational complexity.

Methodology:

Performs centroid-based clustering of categorical vectors or matrices that encode mutation presence, absence, and uncertainty to reduce dataset size for downstream phylogenetic inference.